# Current key AI and technology people

**Research snapshot: 24 October 2026 (UTC).** This is a global, source-linked guide to people who build advanced AI, determine its technological foundations, scrutinize its effects, or govern its development and use. It is a starting point for research, not a ranking, an endorsement, or a contact directory.

## Contents

- [Scope and how to read this guide](#scope-and-how-to-read-this-guide)
- [Frontier-lab leadership and key personnel](#frontier-lab-leadership-and-key-personnel)
- [Emerging labs, world models, robotics, and AI for science](#emerging-labs-world-models-robotics-and-ai-for-science)
- [Compute, platforms, open infrastructure, and capital](#compute-platforms-open-infrastructure-and-capital)
- [Researchers, evaluation, and public-interest voices](#researchers-evaluation-and-public-interest-voices)
- [Policymakers and government officials](#policymakers-and-government-officials)
- [How influence connects across these groups](#how-influence-connects-across-these-groups)
- [Maintenance and open verification questions](#maintenance-and-open-verification-questions)
- [Sources](#sources)

## Scope and how to read this guide

**Coverage.** The core lab roster includes OpenAI, Anthropic, Google DeepMind, Meta, Microsoft AI, Amazon, xAI/Grok, Mistral, Cohere, DeepSeek, Alibaba/Qwen, ByteDance/Seed, Tencent/Hunyuan, Baidu/ERNIE, Z.ai/Zhipu, Moonshot/Kimi, MiniMax, StepFun, and Xiaomi/MiMo. An adjacent roster covers new general-intelligence ventures, world-model and robotics companies, scientific-discovery labs, and important regional model developers. “Frontier” is used as a practical research category, not a claim that every organization has equivalent capabilities or a certified place on a benchmark leaderboard.

**What “leaders” means here.** For each core organization, identify its company or division head and publicly identifiable scientific/technical leadership, then selected people with substantial influence over safety, products, governance, or deployment. This does not attempt to enumerate every executive, board member, researcher, AI startup, or national official. The technology and policy landscape has no fixed, universally accepted boundary; omissions should be treated as research gaps, not judgments that a person is unimportant.

**Evidence versus assessment.** The role/affiliation column records a publicly documented position. The “why follow” column is an assessment of the decisions or research questions that make that position consequential; it does not imply that one person independently created a model or controls every organizational decision. Team achievements are not individual achievements.

**Freshness and uncertainty.** Official directories and first-party announcements take precedence where available; dated reporting supplements them. An old appointment announcement establishes an appointment, not perpetual tenure. Dates in the source notes identify that distinction. “Announced move,” “reported,” and “pending” are deliberate qualifications: they must not be silently converted into a completed appointment, an official title, or a closed transaction. A snapshot date is not a guarantee that every organization has disclosed every subsequent personnel change.

**Reading order.** For model-development decisions, start with the lab sections. For constraints on scale, read compute and platforms. For competing assessments of benefits and harms, read researchers and public-interest voices. For binding powers, distinguish lawmakers, regulators, procurement officials, and political executives from scientific advisers and international conveners.

## Frontier-lab leadership and key personnel

### OpenAI — research, deployment, and nonprofit control

OpenAI's published structure distinguishes the OpenAI Foundation from OpenAI Group PBC. The Foundation controls the Group and appoints its board. Operational leadership, corporate board membership, and the Foundation's safety-governance authority are therefore different kinds of influence.[^openai-structure]

| Person | Public role or affiliation | Why follow their work or decisions | Evidence |
| --- | --- | --- | --- |
| **Sam Altman** | OpenAI CEO and cofounder; Foundation director. | Company strategy, research funding, infrastructure partnerships, and the public case for frontier AI deployment. | [^openai-structure][^openai-reset] |
| **Greg Brockman** | OpenAI president and cofounder; expanded product and business responsibilities in 2026. | Connects model research, infrastructure, product execution, and enterprise adoption; do not rely on an older org chart that places these responsibilities elsewhere. | [^openai-reset] |
| **Jakub Pachocki** | Chief scientist. | Scientific direction of advanced models and the implications of accelerating capabilities; a central research counterpart to Mark Chen. | [^openai-science][^madry-return] |
| **Mark Chen** | Chief research officer. | Research leadership spanning capabilities and safety, and translation of scientific progress into products. | [^openai-research] |
| **Sarah Friar** | Chief financial officer. | Financing and resource allocation determine which large-scale training and infrastructure plans can actually proceed. | [^openai-finance][^openai-reset] |
| **Dali Rajic** | Chief revenue officer, appointed August 2026. | Enterprise deployment, the revenue organization, and commercial feedback into the product roadmap. | [^openai-rajic] |
| **Lilian Weng** | Returned to OpenAI in July 2026 to lead a research-acceleration team. | Work on using AI to improve AI research; previously led safety work. She should no longer be listed as current Thinking Machines leadership. | [^weng-return] |
| **Noam Shazeer** | Announced a move from Google to OpenAI in June 2026; precise new title/start date not established by the cited announcement. | Transformer and language-model architecture expertise; follow his research rather than inventing an OpenAI executive title. | [^shazeer-move] |
| **Aleksander Mądry** | Announced a return to work with Pachocki on safety/security on 19 October 2026. | Robustness, security, the pace of AI progress, and associated strategy. “Former Preparedness head” is not evidence that he has resumed that title. | [^madry-return] |
| **Bret Taylor** | Chair of the OpenAI Foundation board; also a Group director. | Governance, leadership accountability, and the nonprofit/commercial relationship. | [^openai-structure] |
| **Zico Kolter** | Foundation director and chair of its Safety and Security Committee; not a voting Group director. | Formal safety oversight, distinct from operational management or an external advisory role. | [^openai-structure] |
| **Paul Christiano** | Foundation and Group director. | Alignment-research expertise at board level; distinguish this board position from his earlier government and research appointments. | [^openai-structure] |
| **Paul Nakasone** | Foundation and Group director. | National-security and cybersecurity expertise in corporate governance. | [^openai-structure] |

**Roster correction:** Denise Dresser, Brad Lightcap, and Fidji Simo appear in many older leadership lists. August reporting describes their departures from operating roles, with Simo remaining an adviser. Do not carry their older operating titles forward without a newer source.[^openai-reset]

### Anthropic — Claude, alignment, interpretability, and public benefit

| Person | Public role or affiliation | Why follow their work or decisions | Evidence |
| --- | --- | --- | --- |
| **Dario Amodei** | Cofounder and CEO. | Frontier-model strategy, scaling, deployment choices, and the company's approach to benefits and risks. | [^anthropic-leadership] |
| **Daniela Amodei** | Cofounder, president, and board chair. | Company operations and governance; her chair role should not be confused with an independent outside directorship. | [^anthropic-leadership] |
| **Jared Kaplan** | Cofounder and chief science officer. | Scientific strategy, model scaling, and evaluation of future capabilities. | [^anthropic-leadership] |
| **Tom Brown** | Cofounder and chief compute officer. | The compute and systems resources that make frontier research feasible. | [^anthropic-leadership] |
| **Sam McCandlish** | Cofounder and chief architect. | Architecture and technical direction; the current directory does not identify him as CTO. | [^anthropic-leadership] |
| **Chris Olah** | Cofounder; interpretability research leader. | Understanding model internals and whether mechanistic explanations can improve auditing and safety. | [^anthropic-leadership][^anthropic-alignment] |
| **Jack Clark** | Cofounder; head of public benefit and the Anthropic Institute. | Connects research, societal impacts, public-benefit commitments, and policy debates. | [^anthropic-leadership] |
| **Rahul Patil** | Chief technology officer. | Engineering leadership and the reliability and infrastructure requirements of growing deployment. | [^anthropic-leadership] |
| **Ami Vora** | Chief product officer. | How Claude's capabilities become products and how user needs shape deployment. | [^anthropic-leadership] |
| **Krishna Rao** | Chief financial officer. | Capital allocation and financing for the company's compute-intensive research and deployment. | [^anthropic-leadership] |
| **Mariano-Florentino “Tino” Cuéllar** | Head of global affairs. | The company's engagement with governments and institutions around AI governance and public benefit. | [^anthropic-leadership] |
| **Jan Leike** | Leads the Alignment Science team. | Scalable oversight, weak-to-strong generalization, and aligning systems whose work humans cannot easily evaluate. | [^leike][^anthropic-alignment] |
| **Amanda Askell** | Researcher working on Claude's character and alignment. | How behavioral principles and reasoning about values are translated into model behavior. | [^anthropic-alignment] |
| **Boris Cherny** | Creator of Claude Code. | Agentic software development and the design of tools that can change the software-production workflow. | [^claude-code] |
| **Andrej Karpathy** | Joined Anthropic's pretraining team in May 2026. | Model training, AI-assisted research, and unusually influential technical education; no longer best described solely as an independent educator. | [^karpathy-move] |
| **John Jumper** | Announced a move from Google DeepMind to Anthropic in June 2026, after a break; exact new title not specified. | AlphaFold research leadership and AI-for-science expertise. Do not infer that the announcement alone established his start date or remit. | [^jumper-move] |

### Google DeepMind and Alphabet

**Important August 2026 change:** Koray Kavukcuoglu now leads Google DeepMind as senior vice president; Demis Hassabis became its chair and Alphabet's chief scientist. Jeff Dean and colleagues left to establish Discovery Loop. An older list headed “DeepMind CEO Demis Hassabis; Google chief scientist Jeff Dean” is not a current operating roster.[^google-reorg][^discovery-loop]

| Person | Public role or affiliation | Why follow their work or decisions | Evidence |
| --- | --- | --- | --- |
| **Sundar Pichai** | CEO of Google and Alphabet. | Allocation across frontier models, search, cloud, applications, and infrastructure. | [^google-reorg] |
| **Koray Kavukcuoglu** | Senior vice president leading Google DeepMind; Google's chief AI architect. | Day-to-day scientific/organizational leadership and integration of Gemini across Google. | [^google-reorg] |
| **Demis Hassabis** | Google DeepMind chair and Alphabet chief scientist; continues leading Isomorphic Labs. | Long-term scientific direction and AI-driven drug discovery, rather than the former daily GDM CEO remit. | [^google-reorg] |
| **Pushmeet Kohli** | GDM research vice president leading AI for Science and Strategic Initiatives; Google Cloud chief scientist. | Scientific applications and the connection between research breakthroughs and cloud customers. | [^kohli] |
| **Anca Dragan** | Vice president, AI Safety & Behavior. | Safety, alignment, behavior, and evaluation of the Gemini model family. | [^gemini-safety] |
| **Raluca Ada Popa** | Head of Gemini security. | Security engineering and defenses around increasingly capable models; distinguish this remit from general model-safety leadership. | [^gemini-safety] |

### Meta — Superintelligence Labs and FAIR

The cited leadership transcript and reorganization reporting establish the named research roles in 2025; they are not a complete fresh 2026 internal org chart. Meta's company-level leadership directory is the preferred starting point for later updates.[^meta-leadership][^meta-research]

| Person | Public role or affiliation | Why follow their work or decisions | Evidence |
| --- | --- | --- | --- |
| **Mark Zuckerberg** | Founder, chairman, and CEO of Meta. | Compute spending, research priorities, release strategy, and distribution through consumer platforms. | [^meta-leadership] |
| **Alexandr Wang** | Chief AI officer; leads Meta Superintelligence Labs. | Research organization, talent recruitment, and the route from frontier models to Meta's products. | [^meta-leadership][^meta-research] |
| **Nat Friedman** | Leads AI products and applied research within MSL. | Product deployment and developer-facing adoption; not interchangeable with the chief scientist role. | [^meta-research] |
| **Shengjia Zhao** | MSL chief scientist, as documented in the cited company transcript. | Scientific direction of the frontier-model effort. | [^meta-research] |
| **Rob Fergus** | Head of FAIR, as documented in the cited company transcript. | Long-horizon research and its relationship to product-oriented frontier-model development. | [^meta-research] |

Yann LeCun is listed under AMI Labs below, not as Meta's current chief AI scientist.[^ami]

### Microsoft AI and Amazon model development

| Person | Public role or affiliation | Why follow their work or decisions | Evidence |
| --- | --- | --- | --- |
| **Satya Nadella** | Microsoft chairman and CEO. | Research versus partner-model strategy, Azure investment, and distribution through productivity and developer tools. | [^microsoft-officers][^microsoft-copilot] |
| **Mustafa Suleyman** | CEO of Microsoft AI; focused on frontier-model/superintelligence work following the March 2026 reorganization. | Microsoft's independent model capabilities; his remit should not be conflated with leadership of every Copilot product. | [^microsoft-copilot] |
| **Karén Simonyan** | Microsoft AI chief scientist. | Scientific direction of Microsoft's own advanced models and training approaches. | [^microsoft-ai-science] |
| **Ali Farhadi** | Microsoft AI researcher; retains a University of Washington faculty affiliation. | Open-model, visual, and multimodal research experience brought to Microsoft's model effort after leaving the Ai2 CEO role. | [^microsoft-ai2] |
| **Hannaneh “Hanna” Hajishirzi** | Microsoft AI researcher; retains a University of Washington faculty affiliation. | Language-model research and experience developing open research artifacts; no longer assigned her former Ai2 operating role here. | [^microsoft-ai2] |
| **Andy Jassy** | Amazon president and CEO. | Investment priorities linking models, custom chips, AWS capacity, and consumer/enterprise applications. | [^amazon-models][^amazon-team] |
| **Peter DeSantis** | Amazon SVP for foundational AI models, custom silicon, and quantum computing. | Nova and other model research alongside Trainium/Graviton/Nitro: an especially direct link between model and infrastructure strategy. | [^amazon-models][^amazon-team] |
| **Pieter Abbeel** | Amazon frontier-model research lead, appointed in the December 2025 reorganization and confirmed in July 2026 reporting. | Scientific direction of the base-model team and the transfer of robotics/reinforcement-learning expertise into foundation models. | [^amazon-models][^amazon-lab-closure] |

Amazon's December 2025 reorganization replaced an older Rohit Prasad-led structure. Subsequent reporting describes the July 2026 closure of the San Francisco AGI lab; do not assume every leader in the earlier Adept/AGI-lab roster retains that remit.[^amazon-models][^amazon-lab-closure]

### xAI/Grok — within SpaceX

Use the durable research/product label **xAI/Grok (within SpaceX)** when following this organization across changing corporate branding. The cited 2026 reporting describes SpaceX ownership and Michael Nicolls's operating leadership; it does not support treating all original xAI cofounders as current executives.[^xai]

| Person | Public role or affiliation | Why follow their work or decisions | Evidence |
| --- | --- | --- | --- |
| **Elon Musk** | xAI founder; SpaceX CEO. | Capital, computing infrastructure, Grok's direction, and integration with the wider SpaceX/X ecosystem. | [^xai] |
| **Michael Nicolls** | President of the xAI/SpaceX AI organization, appointed in spring 2026. | Engineering execution, model-development organization, and deployment priorities. | [^xai] |

### Mistral AI — France and Europe

| Person | Public role or affiliation | Why follow their work or decisions | Evidence |
| --- | --- | --- | --- |
| **Arthur Mensch** | Cofounder and CEO. | Frontier models, enterprise/government deployment, and Europe's capacity to develop its own AI stack. | [^mistral] |
| **Guillaume Lample** | Cofounder and chief science officer. | Model research, training efficiency, and scientific choices behind Mistral's releases. | [^mistral] |
| **Timothée Lacroix** | Cofounder and CTO. | Engineering and systems execution connecting model research to usable infrastructure. | [^mistral] |

### Cohere — Canada and enterprise AI

| Person | Public role or affiliation | Why follow their work or decisions | Evidence |
| --- | --- | --- | --- |
| **Aidan Gomez** | Cofounder and CEO. | Enterprise models, retrieval, private deployment, and the business case for an independent Canadian AI developer. | [^cohere] |
| **Nick Frosst** | Cofounder. | Founding scientific perspective and the company's enterprise-focused model strategy; no additional current executive title is assumed. | [^cohere] |
| **Ivan Zhang** | Cofounder. | Founding engineering/product perspective; distinguish the documented cofounder designation from an unverified current functional title. | [^cohere] |
| **Joelle Pineau** | Chief AI officer. | Scientific leadership and enterprise AI research following her move from Meta. | [^cohere] |
| **Phil Blunsom** | Chief technology officer. | Technical strategy and implementation; the current directory identifies Blunsom, not Frosst or Zhang, as CTO. | [^cohere] |

### China — major platform model developers

**Names and boundaries:** Chinese names generally appear family-name first below; familiar English names are included where useful. The division head, platform CEO, and individual model contributor may have very different authority. Public internal reporting lines can be less complete than company-level leadership disclosures.

| Organization | Person | Public role or affiliation | Why follow their work or decisions | Evidence |
| --- | --- | --- | --- | --- |
| DeepSeek | **Liang Wenfeng** | Founder and CEO. | Research priorities, efficient frontier-model development, and the company's open-weight strategy. | [^deepseek] |
| DeepSeek | **Chen Deli / Victor Chen** | Senior researcher and public representative. | A visible technical voice on DeepSeek and AI's societal effects; the cited reporting does **not** establish a chief scientist title. | [^deepseek] |
| Alibaba / Qwen | **Eddie Wu / Wu Yongming** | Alibaba Group CEO; leads its reorganized AI effort. | Group-wide allocation of talent, compute, and commercialization around Qwen and associated platforms. | [^alibaba-april][^alibaba-june] |
| Alibaba / Qwen | **Zhou Jingren / Jingren Zhou** | Appointed Alibaba chief scientist and head of the AI Future Research Institute in June 2026. | Longstanding Qwen/Tongyi architect; the June restructuring supersedes the April chief-AI-architect/Cloud-CTO descriptions. | [^alibaba-june] |
| Alibaba | **Wu Zeming** | Group CTO. | Business technology and AI inference platforms, distinct from the foundation-model research role. | [^alibaba-april] |
| Alibaba Cloud | **Li Feifei** | Cloud CTO, appointed April 2026. | Cloud technology and AI infrastructure. This is **not** World Labs cofounder Fei-Fei Li. | [^alibaba-april] |
| ByteDance / Seed | **Liang Rubo** | ByteDance CEO. | Investment and deployment across Seed/Doubao and the company's larger platform businesses. | [^bytedance] |
| ByteDance / Seed | **Wu Yonghui** | Head of foundational research for Seed. | Model algorithms, infrastructure, multimodal research, and longer-term research direction; do not invent a formal “chief scientist” title. | [^bytedance] |
| Tencent / Hunyuan | **Pony Ma / Ma Huateng** | Tencent chairman and CEO. | AI investment and distribution through Tencent's consumer, gaming, and cloud ecosystem. | [^tencent-leadership] |
| Tencent / Hunyuan | **Martin Lau / Liu Chiping** | Tencent president. | Company-level strategy and the management interface for the AI scientific effort. | [^tencent-leadership][^tencent-ai] |
| Tencent / Hunyuan | **Yao Shunyu / Shunyu Yao** | Chief AI scientist. | Hunyuan research and the integration of model and infrastructure work following Tencent's AI reorganization. | [^tencent-ai] |
| Baidu / ERNIE | **Robin Li / Li Yanhong** | Cofounder, chairman, and CEO. | ERNIE's strategy and its connection to search, cloud services, and autonomous driving. | [^baidu] |
| Baidu / ERNIE | **Wang Haifeng / Haifeng Wang** | CTO. | AI technology strategy and deep-learning infrastructure; do not assume that the title alone establishes every current model-team reporting line. | [^baidu] |
| Xiaomi / MiMo | **Lei Jun** | Founder, chairman, and CEO. | Model investment and distribution through phones, connected devices, and vehicles. | [^xiaomi] |
| Xiaomi / MiMo | **Luo Fuli / Fuli Luo** | Leads MiMo. | Model training and the connection between reasoning models and physical-device applications. She should not be listed as current DeepSeek personnel. | [^xiaomi] |

### China — independent model companies

| Organization | Person | Public role or affiliation | Why follow their work or decisions | Evidence |
| --- | --- | --- | --- | --- |
| Z.ai / Zhipu | **Zhang Peng** | CEO. | Commercial and organizational direction of the GLM model developer. | [^zhipu] |
| Z.ai / Zhipu | **Liu Debing** | Chairman. | Corporate governance and the allocation of resources around model development. | [^zhipu] |
| Z.ai / Zhipu | **Tang Jie** | Cofounder and chief scientist; Tsinghua professor. | The research lineage linking knowledge systems and foundation models. | [^zhipu] |
| Moonshot / Kimi | **Yang Zhilin / Zhilin Yang** | Cofounder and CEO. | Long-context/reasoning models, open-weight releases, and Kimi's consumer/developer strategy. | [^moonshot] |
| MiniMax | **Yan Junjie** | Founder, chairman, CEO, and CTO. | Direct scientific and business leadership across text, speech, and visual models. | [^minimax] |
| MiniMax | **Yun Yeyi** | Executive director and president; COO until March 2026. | Product strategy, commercialization, and operating execution. The dated May filing supersedes the older management-page COO title. | [^minimax] |
| MiniMax | **Zhao Pengyu** | Executive director; large-language-model research and engineering leader. | Scientific/engineering execution of MiniMax's language-model program. | [^minimax] |
| MiniMax | **Zhou Yucong** | Executive director; visual-model research and engineering leader. | The visual-generation side of MiniMax's multimodal portfolio. | [^minimax] |
| StepFun | **Jiang Daxin** | Founder and CEO. | General-purpose model development and its deployment in consumer and industrial systems. | [^stepfun] |
| StepFun | **Yin Qi** | Chairman, appointed January 2026. | Company strategy and links between foundation models, intelligent vehicles, and devices. | [^stepfun] |
| StepFun | **Zhang Xiangyu** | Chief scientist. | Scientific direction of the model effort, including the contribution of computer-vision expertise. | [^stepfun] |
| StepFun | **Zhu Yibo** | CTO. | Training/inference systems and the engineering scale required by the research program. | [^stepfun] |

## Emerging labs, world models, robotics, and AI for science

These organizations are included for their research direction, talent, regional significance, or infrastructure commitments—not because a founding announcement proves a capability breakthrough. Claims about achieving superintelligence remain the organizations' ambitions, not findings of this guide.

### New general-intelligence and research-automation ventures

| Organization | Person | Public role or affiliation | Why follow their work or decisions | Evidence |
| --- | --- | --- | --- | --- |
| Safe Superintelligence (SSI) | **Ilya Sutskever** | Cofounder and CEO; leads the technical team. | A focused attempt to build safe superintelligence outside a conventional near-term product roadmap. | [^ssi] |
| SSI | **Daniel Levy** | Cofounder and president. | Executive leadership alongside Sutskever; Daniel Gross is no longer part of SSI in the cited company update. | [^ssi] |
| Thinking Machines Lab | **Mira Murati** | Founder and CEO. | Human–AI collaboration, customizable models, and research/product co-design. | [^thinking-machines] |
| Thinking Machines Lab | **John Schulman** | Cofounder and chief scientist. | Reinforcement learning, post-training, and making capable systems more understandable and adaptable. | [^thinking-machines] |
| Thinking Machines Lab | **Soumith Chintala** | CTO since January 2026; PyTorch cocreator. | Research infrastructure and the developer ecosystem needed for customizable AI. | [^thinking-machines] |
| Discovery Loop | **Jeff Dean** | Cofounder; departed Google in August 2026. | Research automation informed by decades of large-scale computing and machine-learning systems work. | [^discovery-loop] |
| Discovery Loop | **Sanjay Ghemawat** | Cofounder; departed Google. | Distributed systems and the infrastructure needed for large numbers of automated experiments. | [^discovery-loop] |
| Discovery Loop | **Oriol Vinyals** | Cofounder; formerly a GDM research vice president. | Sequence modeling, reinforcement learning, and research automation. | [^discovery-loop] |
| Discovery Loop | **Quoc Le** | Cofounder; formerly a Google Brain cofounder. | Representation learning and automating the machine-learning research process. | [^discovery-loop] |
| Ineffable Intelligence | **David Silver** | Founder. | Reinforcement learning from experience as an alternative or complement to learning primarily from human-produced data. The cited mission is a research ambition, not a demonstrated outcome. | [^ineffable] |
| Recursive | **Richard Socher** | Cofounder; publicly reported company leader. | Self-improving research systems and an attempt to automate the scientific method. | [^recursive] |
| Recursive | **Tim Rocktäschel** | Cofounder and researcher. | Open-ended learning and automated discovery; no unverified current functional executive title is assigned here. | [^recursive] |
| Periodic Labs | **Liam Fedus** | Cofounder. | Combining AI models with physical experimentation, rather than evaluating scientific progress only through text benchmarks. | [^periodic] |
| Periodic Labs | **Ekin Doğuş Çubuk** | Cofounder and co-CEO. | Materials discovery, experimental feedback, and the scientific reliability of autonomous-laboratory systems. | [^periodic] |

Thinking Machines' original founding roster is not its current leadership roster: Barret Zoph and Luke Metz left in January 2026; Lilian Weng returned to OpenAI in July.[^thinking-machines][^weng-return]

### World models and physical intelligence

| Organization | Person | Public role or affiliation | Why follow their work or decisions | Evidence |
| --- | --- | --- | --- | --- |
| AMI Labs | **Yann LeCun** | Cofounder and executive chair. | World models, representation learning, memory, and planning beyond language-only approaches. | [^ami] |
| AMI Labs | **Alexandre “Alex” LeBrun** | Cofounder and CEO. | Translating AMI's research program into a sustainable organization and applications. | [^ami] |
| AMI Labs | **Saining Xie** | Cofounder and chief science officer. | Visual representation learning and the scientific program behind the lab's world-model approach. | [^ami] |
| AMI Labs | **Pascale Fung** | Cofounder; chief research and innovation officer. | Multimodal/human-centered AI and the path from world-model research to real-world applications. | [^ami] |
| AMI Labs | **Mike / Michael Rabbat** | Cofounder; vice president of world models. | The central world-model research agenda. | [^ami] |
| World Labs | **Fei-Fei Li** | Cofounder and CEO; announced future AMD EVP/chief scientist role, conditional on the transaction closing. | Spatial intelligence and computer vision. The September agreement to join AMD is **pending**, not evidence that she already holds the AMD role. | [^world-labs] |
| World Labs | **Justin Johnson** | Cofounder; continuing research/team leadership in the announced AMD combination. | Visual/spatial models and the scientific direction of the World Labs team. | [^world-labs] |
| World Labs | **Ben Mildenhall** | Cofounder; continuing research/team leadership in the announced AMD combination. | Three-dimensional representations and the connection between graphics and generative AI. | [^world-labs] |
| Physical Intelligence (π) | **Karol Hausman** | Cofounder and CEO. | General-purpose robot foundation models and the commercial organization around them. | [^physical-intelligence] |
| Physical Intelligence (π) | **Sergey Levine** | Cofounder; UC Berkeley professor. | Robot learning and reinforcement learning that connects perception to physical action. | [^physical-intelligence] |
| Physical Intelligence (π) | **Chelsea Finn** | Cofounder; Stanford professor. | Generalization, learning from demonstrations, and adaptation across robotic tasks. | [^physical-intelligence] |

### Open scientific-model research

| Organization | Person | Public role or affiliation | Why follow their work or decisions | Evidence |
| --- | --- | --- | --- | --- |
| Allen Institute for AI (Ai2) | **Peter Clark** | Interim CEO and senior research director; co-leads Asta. | Nonprofit open-model research, scientific reasoning, and agentic systems for assisted/automated discovery. The interim designation matters; Ali Farhadi is not its current CEO. | [^ai2] |
| Ai2 | **Bill Hilf** | Board chair. | Governance and institutional continuity during the CEO transition, separate from day-to-day scientific leadership. | [^ai2] |

### Regional model-building capacity

| Organization | Person | Public role or affiliation | Why follow their work or decisions | Evidence |
| --- | --- | --- | --- | --- |
| Sakana AI — Japan | **David Ha** | Cofounder and CEO. | Nature-inspired approaches, model development, and AI research tailored to Japan's needs. | [^sakana] |
| Sakana AI — Japan | **Llion Jones** | Cofounder and CTO; Transformer-paper coauthor. | Model architecture and the technical direction of a significant independent Japanese developer. | [^sakana] |
| Sakana AI — Japan | **Ren Ito** | Cofounder and chairman in the current directory. | Corporate strategy and institutional partnerships; older sources identify him as COO. | [^sakana] |
| Technology Innovation Institute — UAE | **Najwa Aaraj** | CEO of TII. | Falcon development, Arabic-language models, and domestic AI research capability. | [^tii] |
| Advanced Technology Research Council — UAE | **Faisal Al Bannai** | ATRC secretary general and adviser to the UAE president. | Public research priorities and the institutional support behind TII/Falcon. | [^tii] |
| Sarvam AI — India | **Vivek Raghavan** | Cofounder. | Indian-language models and connections to digital public infrastructure; the source establishes cofounder status without requiring a disputed CEO label. | [^sarvam] |
| Sarvam AI — India | **Pratyush Kumar** | Cofounder. | Indian-language/open AI research and the engineering of a domestically developed model stack. | [^sarvam] |

## Compute, platforms, open infrastructure, and capital

Model builders do not control the entire stack. Chip design, fabrication, lithography, cloud capacity, operating systems, developer platforms, distribution, and financing create different sources of leverage. People already profiled above, including Nadella, Pichai, Jassy, Zuckerberg, and Musk, belong in this picture too.

### Semiconductors and the physical computing stack

| Person | Public role or affiliation | Why follow their work or decisions | Evidence |
| --- | --- | --- | --- |
| **Jensen Huang** | NVIDIA founder, president, and CEO. | Accelerators, networking, the CUDA ecosystem, and investments or partnerships that shape access to AI infrastructure. | [^nvidia] |
| **Lisa Su** | AMD chair and CEO. | Accelerator competition, software support, and alternatives to a single-vendor AI compute stack. | [^amd] |
| **C. C. Wei / Wei Che-chia** | TSMC chairman and CEO. | Advanced-node fabrication and packaging capacity: model ambition is constrained by manufactured hardware, not just chip designs. | [^tsmc] |
| **Christophe Fouquet** | ASML president and CEO. | Lithography technology and supply-chain capacity underlying leading-edge semiconductor production. | [^asml] |
| **Hock Tan** | Broadcom president and CEO. | Custom AI accelerators and networking, including the economics of hyperscalers developing their own chips. | [^broadcom] |
| **Lip-Bu Tan** | Intel CEO. | Foundry execution, advanced packaging, and competition in computing and semiconductor manufacturing. | [^intel] |
| **Rene Haas** | Arm CEO; also CEO of SoftBank Group International. | Processor architecture and coordination across a semiconductor/AI portfolio; architecture licensing is distinct from fabrication. | [^arm] |
| **Ren Zhengfei** | Huawei founder; officially listed as director and CEO. | Huawei's long-term strategy across telecommunications, cloud, and domestic computing alternatives. Huawei's rotating-chair system is a separate layer of operational governance. | [^huawei] |

### Cloud capacity, enterprise deployment, devices, and financing

| Person | Public role or affiliation | Why follow their work or decisions | Evidence |
| --- | --- | --- | --- |
| **Matt Garman** | AWS CEO. | Cloud availability, enterprise access to models, custom-chip adoption, and the infrastructure economics of deployment. | [^amazon-team] |
| **Jay Parikh** | Microsoft EVP, CoreAI. | The engineering platform and developer infrastructure through which organizations build AI applications. | [^microsoft-officers] |
| **Jacob Andreou** | Microsoft EVP, Copilot. | Leads the consumer/commercial Copilot experience following the March 2026 reorganization; product authority differs from Suleyman's model-research remit. | [^microsoft-copilot] |
| **Brad Smith** | Microsoft vice chair and president. | Government relations, legal commitments, cybersecurity, and the policy conditions attached to global AI infrastructure. | [^microsoft-officers] |
| **Thomas Kurian** | Google Cloud CEO. | Enterprise AI distribution and cloud partnerships, including access to models from developers other than Google. | [^google-cloud] |
| **Larry Ellison** | Oracle chairman and CTO. | Cloud infrastructure and database strategy; distinguish his strategic influence from the co-CEOs' operating roles. | [^oracle] |
| **Clay Magouyrk** | Oracle CEO, alongside Mike Sicilia. | The cloud-infrastructure side of Oracle's AI expansion and delivery commitments. | [^oracle] |
| **Mike Sicilia** | Oracle CEO, alongside Clay Magouyrk. | Enterprise and industry applications, including how AI agents are embedded into existing business software. | [^oracle] |
| **Safra Catz** | Oracle executive vice chair. | Board-level strategy and continuity; she should no longer be listed as Oracle's sole CEO. | [^oracle] |
| **John Ternus** | Apple CEO since 1 September 2026. | Device integration, on-device AI, and access to a major consumer-computing distribution channel. | [^apple] |
| **Tim Cook** | Apple executive chair. | Board-level continuity and relationships with governments; not Apple's current CEO. | [^apple] |
| **Craig Federighi** | Apple SVP, Software Engineering. | Operating-system integration, user experience, privacy architecture, and the practical deployment of AI on devices. | [^apple] |
| **Johny Srouji** | Apple chief hardware officer. | The silicon and device capabilities that constrain local inference and its cost/energy trade-offs. | [^apple] |
| **Jennifer Newstead** | Apple SVP, General Counsel and Government Affairs. | The legal and governmental interface for platform access, privacy, competition, and deployment. | [^apple] |
| **Michael “Mike” Intrator** | CoreWeave cofounder, chairman, and CEO. | Specialized AI-cloud capacity, financing, and the infrastructure commitments supporting model developers. | [^coreweave] |
| **Masayoshi Son** | SoftBank Group chairman and CEO. | Financing and industrial coordination across AI models, semiconductors, and infrastructure. Announced investment plans are not the same as deployed capacity. | [^softbank] |
| **Peng Xiao** | G42 group CEO. | Sovereign AI infrastructure, international partnerships, and applications across the UAE-based group's businesses. | [^g42] |
| **Tareq Amin** | HUMAIN CEO. | Saudi Arabia's attempt to connect AI infrastructure, cloud, models, and applications in one operating company. | [^humain] |
| **Arvind Krishna** | IBM chairman, president, and CEO. | Enterprise AI, hybrid cloud, and the interaction of AI with quantum-computing and industrial research strategies. | [^ibm] |

### Open models and shared software infrastructure

| Person | Public role or affiliation | Why follow their work or decisions | Evidence |
| --- | --- | --- | --- |
| **Clément Delangue** | Hugging Face cofounder and CEO. | The governance and business model of an important model/dataset distribution platform. | [^hugging-face] |
| **Julien Chaumond** | Hugging Face cofounder and CTO. | Developer infrastructure, model sharing, and deployment tooling. | [^hugging-face] |
| **Thomas Wolf** | Hugging Face cofounder and chief science officer. | Open research, scientific tooling, and the relationship between model releases and reproducibility. | [^hugging-face] |
| **Linus Torvalds** | Linux creator and Linux Foundation fellow; leads kernel integration. | A foundational layer of cloud and device computing, maintained through a different governance model from proprietary AI platforms. | [^linux] |

**Pending transaction:** NVIDIA announced an agreement to acquire Hugging Face on 3 September 2026. The expected closing is in the first half of 2027, subject to approvals; the platform should not be described here as already acquired. NVIDIA's promises about continued openness are company commitments, not an independent guarantee about future governance.[^hugging-face-deal]

## Researchers, evaluation, and public-interest voices

This section deliberately includes competing schools of thought. Technical safety, measured capabilities, civil rights, labor conditions, privacy, and scientific reliability ask different questions. An affiliation is not proof of independence; check funding, board roles, and commercial relationships before treating an assessment as disinterested.

### Scientific direction, learning, and evaluation methods

| Person | Public role or affiliation | Why follow their work or decisions | Evidence |
| --- | --- | --- | --- |
| **Yoshua Bengio** | LawZero co-president and scientific director; Université de Montréal professor; Mila founder and scientific adviser. | Deep learning, alternatives for building safer AI, and synthesis of evidence as chair of the International AI Safety Report. He is not described here using his former Mila scientific-director title. | [^bengio] |
| **Geoffrey Hinton** | University of Toronto emeritus distinguished professor; former Google researcher. | Foundational neural-network research and influential public arguments about advanced AI risks; distinguish research reputation from current operating authority at a lab. | [^hinton] |
| **Richard “Rich” Sutton** | University of Alberta professor and Amii chief scientific adviser; reported cofounder of Oak Lab in July 2026. | Reinforcement learning, continual learning, and the argument for learning from experience. Earlier sources listing Keen Technologies predate his reported departure. | [^sutton] |
| **Stuart Russell** | UC Berkeley computer scientist and director of the Center for Human-Compatible AI. | Human-compatible objectives, uncertainty about human preferences, and the intellectual foundations of AI governance. | [^russell] |
| **Percy Liang** | Stanford computer-science professor; director of the Center for Research on Foundation Models. | Transparent model evaluation, reproducibility, and the distinction between benchmark performance and dependable real-world use. | [^liang] |
| **Yejin Choi** | Stanford computer-science professor and HAI senior fellow; directory records partial leave beginning October 2026. | Common-sense reasoning, smaller models, and the limits of scaling as an answer to reasoning and understanding. | [^choi] |
| **Andrew Ng** | Founder of DeepLearning.AI; AI Fund managing general partner; Stanford adjunct professor. | AI education, practical adoption, and startup formation: influence comes through skills and application choices as well as new model architectures. | [^ng] |
| **Dawn Song** | UC Berkeley professor and co-director of its Responsible Decentralized Intelligence center. | Security, privacy, trustworthy agents, and the relationship between decentralization and AI control. | [^song] |
| **Vukosi Marivate** | University of Pretoria professor and AfriDSAI director; Lelapa AI cofounder; Masakhane researcher. | African-language NLP, locally grounded datasets, and research capacity outside the dominant US/China centers. | [^marivate] |
| **Pelonomi Moiloa** | Lelapa AI cofounder and CEO. | Resource-efficient language and speech systems designed around African contexts and languages, rather than assuming that English-first products transfer automatically. | [^lelapa] |
| **Jade Abbott** | Lelapa AI cofounder and CTO. | Low-resource language technologies and the practical infrastructure needed for African-language AI; the current team directory supersedes older COO descriptions. | [^lelapa] |
| **Benjamin Rosman** | Lelapa AI chief scientist; University of the Witwatersrand professor and MIND Institute director. | Robot learning, reinforcement learning, and African research-community capacity, including the Deep Learning Indaba; also chairs South Africa's AI-policy advisory panel. | [^lelapa] |

### Frontier-risk assessment and independent scrutiny

| Person | Public role or affiliation | Why follow their work or decisions | Evidence |
| --- | --- | --- | --- |
| **Beth Barnes** | METR founder and CEO. | Measuring autonomous task completion and whether capabilities warrant stronger safeguards or changes in deployment policy. | [^metr] |
| **Chris Painter** | METR president. | Institutional execution and the connection between evaluations, governance proposals, and public decision-making. | [^metr] |
| **Hjalmar Wijk** | METR chief scientist. | Scientific interpretation of capability measurements and their limits. | [^metr] |
| **Nate Rush** | METR CTO. | Evaluation infrastructure and the engineering required to make capability measurements reliable and reproducible. | [^metr] |
| **Dan Hendrycks** | Center for AI Safety executive and research director; also works on safety at Meta. | Benchmarks, technical safety, and risk-policy advocacy. His industry work matters when assessing institutional independence. | [^hendrycks] |
| **Arvind Narayanan** | Princeton computer-science professor and Center for Information Technology Policy director. | AI-agent evaluation, privacy, and careful separation of demonstrated capabilities from commercial hype. | [^narayanan] |

METR's own disclosures distinguish donations from model-access arrangements and free tokens. Read those disclosures alongside its reports: a nonprofit label alone does not settle every conflict-of-interest question.[^metr]

### Rights, labor, privacy, and the political economy of technology

| Person | Public role or affiliation | Why follow their work or decisions | Evidence |
| --- | --- | --- | --- |
| **Timnit Gebru** | Founder and executive director of the Distributed AI Research Institute (DAIR). | Community-centered AI research, discrimination, data practices, and who bears the costs of technological deployment. | [^dair] |
| **Alex Hanna** | DAIR director of research. | Data labor, social consequences of AI, and critiques of claims that replacement of human work is inevitable. | [^dair] |
| **Milagros Miceli** | DAIR research lead. | Data work and the labor, power, and working conditions behind model training and evaluation. | [^dair] |
| **Joy Buolamwini** | Algorithmic Justice League founder. | Algorithmic discrimination, biometric surveillance, and public accountability for automated systems. | [^buolamwini] |
| **Margaret Mitchell** | Hugging Face chief ethics scientist. | Model documentation, evaluation, responsible release practices, and accountability in widely shared AI systems. | [^mitchell] |
| **Rumman Chowdhury** | Humane Intelligence PBC founder and CEO. | Applied AI evaluation, red-teaming, and accountability. Distinguish the public-benefit corporation from the separately organized Humane Intelligence nonprofit. | [^chowdhury] |
| **Emily M. Bender** | University of Washington linguistics professor and Computational Linguistics Laboratory director. | The difference between language generation and understanding, documentation of language data, and scrutiny of anthropomorphic claims. | [^bender] |
| **Kate Crawford** | USC professor; leads Knowing Machines. | The material, environmental, labor, and political foundations of AI, rather than treating it as immaterial software alone. | [^crawford] |
| **Meredith Whittaker** | Signal president; AI Now cofounder. | Encryption, surveillance business models, and the privacy/security consequences of highly privileged AI assistants. | [^whittaker] |
| **Amba Kak** | AI Now Institute co-executive director. | Concentrated corporate power, competition, and public-interest technology policy. | [^ai-now] |
| **Sarah Myers West** | AI Now Institute co-executive director. | Political economy, surveillance, labor, and how institutions can preserve meaningful public choice over AI's direction. | [^ai-now] |

These people do not represent a single consensus. Read their actual arguments and evidence alongside frontier-lab claims; do not reduce every disagreement to an “AI optimist versus doomer” binary.

## Policymakers and government officials

**Authority is not uniform.** Political executives set priorities and budgets; legislatures pass laws; regulators enforce particular statutes; research institutes supply evidence; international organizations convene and coordinate. An adviser or commission member does not automatically have binding regulatory authority. This is a map of consequential offices and people, not legal advice or a complete inventory of every jurisdiction's AI rules.

### United States — executive coordination, standards, security, and markets

| Person | Public role or affiliation | Relevant authority or influence | Evidence |
| --- | --- | --- | --- |
| **Donald Trump** | US president. | Executive-branch priorities, appointments, procurement, trade policy, and the administration's AI strategy, within statutory and constitutional limits. | [^us-executive][^us-pcast] |
| **JD Vance** | US vice president. | Senior political advocacy and international representation on AI and technology; not a standalone AI regulator. | [^us-executive] |
| **Jay Clayton** | Director of National Intelligence; chair of the White House AI task force established in October 2026. | Cross-government coordination of security, economic, and societal AI concerns under the new task-force structure. | [^us-taskforce] |
| **Michael Kratsios** | White House OSTP director and presidential science/technology adviser; PCAST co-chair. | Scientific strategy and interagency policy coordination; distinguish this standing office from external advisory membership. | [^us-pcast][^ostp] |
| **David Sacks** | PCAST co-chair. | External technology advice and political influence. His special-government-employee stint as AI/crypto adviser ended in March 2026; “current AI czar” is an outdated description. | [^us-pcast][^sacks-transition] |
| **Sean Cairncross** | National cyber director. | Cybersecurity coordination, including the intersection between advanced AI and national cyber resilience. | [^us-taskforce] |
| **Emil Michael** | Under secretary of defense for research and engineering; AI task-force vice-chair. | Defense technology, security assessment, and military adoption; this is distinct from civilian consumer-protection authority. | [^us-taskforce] |
| **Scott Kupor** | Office of Personnel Management director; AI task-force vice-chair. | Federal workforce planning and the government's organizational response to AI. | [^us-taskforce] |
| **Andrew Ferguson** | FTC chair; AI task-force vice-chair. | Consumer-protection and competition enforcement, including coordination concerning AI-related harms. | [^ftc][^us-taskforce] |
| **Howard Lutnick** | Secretary of commerce. | Semiconductor and industrial strategy and the department containing BIS and NIST. | [^commerce] |
| **Jeffrey Kessler** | Under secretary of commerce for industry and security; heads BIS. | Export controls and access to advanced computing technology, subject to administration policy and applicable law. | [^commerce] |
| **Arvind Raman** | NIST director; announced as acting CAISI director after Chris Fall's July 2026 departure. | Standards, measurement, testing, and government technical understanding of AI capabilities and security. The CAISI role is acting, not a confirmed permanent appointment. | [^nist][^caisi] |
| **Brendan Carr** | FCC chair. | Communications networks and infrastructure, and the agency's assigned work on AI reporting/disclosure and federal–state policy questions. | [^fcc] |
| **Darío Gil** | Department of Energy under secretary for science; leads the Genesis Mission. | National-laboratory research, public scientific computing, and AI-for-science infrastructure and collaboration. | [^doe] |

The October task force is a recent coordination structure, not proof that earlier agencies lost their statutory responsibilities. Read its reported membership and priorities separately from enacted law.[^us-taskforce]

### United States — legislators and state-level policy

| Person | Public role or affiliation | Relevant authority or influence | Evidence |
| --- | --- | --- | --- |
| **Ted Cruz** | US senator; Senate Commerce, Science, and Transportation Committee chair. | Legislative agenda and oversight covering technology, communications, and AI-related commerce. | [^senate-commerce] |
| **Maria Cantwell** | US senator; ranking member of Senate Commerce. | The opposition-party leadership counterpart for technology legislation and oversight. | [^senate-commerce] |
| **Jay Obernolte** | US representative; former chair of the House's bipartisan AI task force. | Technical expertise and congressional AI proposals. The task force's 2024 report is a historical contribution, not evidence of a permanent regulator or continuing task-force office. | [^house-ai] |
| **Ted Lieu** | US representative; former co-chair of that House AI task force. | Legislative proposals concerning AI accountability and government use; distinguish proposed bills from enacted law. | [^house-ai] |
| **Gavin Newsom** | Governor of California. | Signing/veto decisions and executive implementation in a jurisdiction hosting many major AI developers. | [^california] |
| **Scott Wiener** | California state senator; author of SB 53. | Frontier-AI transparency and safety legislation, and the continuing debate over state and federal responsibilities. | [^california] |

### United Kingdom, European Union, France, Germany, and Canada

| Jurisdiction | Person | Public role or affiliation | Relevant authority or influence | Evidence |
| --- | --- | --- | --- | --- |
| UK | **Andy Burnham** | Prime minister since 20 July 2026. | Government priorities, public investment, international AI diplomacy, and preparation for AI's effects on work and society. | [^uk-pm] |
| UK | **Kanishka Narayan** | Minister of State for AI, jointly in the Cabinet Office and Department for Business, Innovation, Science and Trade. | AI policy, infrastructure, adoption, sovereign capability, and oversight of the AI Security Institute portfolio. His July appointment supersedes the earlier DSIT title. | [^uk-ai-minister] |
| UK | **Ian Hogarth** | AI Security Institute chair. | Government access to frontier-model evidence and the institute's strategic direction; not the same office as AI minister. | [^aisi] |
| UK | **Chris Summerfield** | AISI research director; Oxford researcher. | Technical assessments of model capability and risk, linking government evidence to the research community. | [^aisi] |
| UK | **Jade Leung** | AISI vice-chair and AI Taskforce security adviser following the September transition. | Strategic AI-security advice. She stepped back from the full-time AISI CTO and prime-ministerial AI-adviser roles at the end of September. | [^leung-transition] |
| EU | **Ursula von der Leyen** | European Commission president. | Commission-wide priorities, investment initiatives, and the political direction of EU AI policy; legislation also requires the other EU institutions. | [^eu-college] |
| EU | **Henna Virkkunen** | Commission EVP for Tech Sovereignty, Security and Democracy. | Digital/frontier technology policy, AI adoption, infrastructure, and implementation/enforcement strategy. | [^virkkunen] |
| EU | **Lucilla Sioli** | Director of the European AI Office within DG CONNECT. | AI Act implementation, coordination of the AI strategy, technical capacity, and international collaboration. | [^sioli] |
| EU | **Roberto Viola** | DG CONNECT director-general, including in the October 2026 organization chart. | The administrative machinery behind digital policy, research programs, and AI infrastructure initiatives. | [^dg-connect] |
| EU | **Teresa Ribera** | Commission EVP for Clean, Just and Competitive Transition; responsible for competition policy. | Competition enforcement, mergers, and market structure, distinct from the AI Office's technical/regulatory remit. | [^ribera] |
| France | **Emmanuel Macron** | President. | National research/industrial priorities, technological sovereignty, and security framing of advanced AI. | [^france] |
| Germany | **Karsten Wildberger** | Federal minister for digital transformation and government modernization. | Digital public infrastructure, state adoption, and sovereign AI partnerships; coordination with other ministries remains important. | [^germany] |
| Canada | **Mark Carney** | Prime minister. | National AI strategy, public investment, and cross-government coordination. | [^canada-strategy] |
| Canada | **Evan Solomon** | Minister of Artificial Intelligence and Digital Innovation. | AI adoption, commercialization, digital sovereignty, and international AI-security partnerships. | [^canada-minister] |

### China and India

| Jurisdiction | Person | Public role or affiliation | Relevant authority or influence | Evidence |
| --- | --- | --- | --- | --- |
| China | **Xi Jinping** | CPC general secretary and state president. | Top-level political direction of AI, technological self-reliance, and the relationship between technological and state power. | [^china-state] |
| China | **Li Qiang** | Premier of the State Council. | Economic implementation, industrial strategy, and international AI-governance initiatives. | [^china-state] |
| China | **Ding Xuexiang** | Vice premier and Politburo Standing Committee member. | Coordination of technology and digital-economy priorities, including computing, talent, and AI development. | [^china-ding] |
| China | **Li Lecheng** | Minister of Industry and Information Technology. | Industrial capacity, communications infrastructure, and implementation of rules affecting AI services. | [^china-cac] |
| China | **Zhuang Rongwen** | Director of the Cyberspace Administration of China. | Internet/data governance and oversight affecting AI services. CAC and MIIT are not interchangeable institutions. | [^china-cac] |
| India | **Narendra Modi** | Prime minister. | The political direction of India's AI strategy and the interface between AI, industrial development, and digital public infrastructure. | [^india-policy] |
| India | **Ashwini Vaishnaw** | Union minister for Electronics and Information Technology, among other portfolios. | AI policy, semiconductor strategy, and the proposed regulatory approach; a forthcoming consultation is not an enacted AI law. | [^india-policy][^india-october] |
| India | **S. Krishnan** | MeitY secretary. | Administrative implementation across AI, data, cybersecurity, and the broader digital economy. | [^india-policy][^krishnan-current] |
| India | **Saurabh Vijay** | UIDAI CEO; additionally appointed CEO of the IndiaAI Mission in June 2026. | Public AI compute, indigenous-model support, datasets, and mission execution. Abhishek Singh's earlier IndiaAI CEO title is not current. | [^indiaai-transition] |

### East/Southeast Asia, Middle East, Africa, and Latin America

| Jurisdiction | Person | Public role or affiliation | Relevant authority or influence | Evidence |
| --- | --- | --- | --- | --- |
| Japan | **Sanae Takaichi** | Prime minister. | National AI/industrial strategy and coordination of the cabinet's digital, security, and implementation portfolios. | [^japan-cabinet] |
| Japan | **Toshiharu Furukawa** | Minister for Digital Transformation; also responsible for AI strategy and AI/digital reform after the September reshuffle. | Government digital transformation and cross-cutting AI strategy, distinct from industry-specific implementation. | [^japan-cabinet] |
| Japan | **Ryosei Akazawa** | Minister of Economy, Trade and Industry; also minister for social implementation of AI transformation. | Industrial adoption, semiconductor policy, and deployment in sectors facing labor shortages. | [^japan-cabinet] |
| South Korea | **Bae Kyung-hoon** | Deputy prime minister and minister of Science and ICT. | AI Basic Act implementation, domestic foundation-model programs, and frontier-computing investment. | [^korea] |
| Singapore | **Josephine Teo** | Minister for Digital Development and Information. | AI governance, public adoption, communications, and cybersecurity in a major regional technology hub. | [^singapore] |
| UAE | **Omar Sultan Al Olama** | Minister of State for AI and director-general of the Prime Minister's Office; chair of the new AI and Data Authority. | Coordination of AI, data, and digital government; federal responsibilities should not be conflated with emirate-level investment bodies. | [^uae] |
| Saudi Arabia | **Abdullah bin Sharaf Alghamdi** | President of the Saudi Data and AI Authority (SDAIA). | National data/AI strategy, governance, capability building, and international coordination. | [^sdaia] |
| Saudi Arabia | **Abdullah Alswaha** | Minister of Communications and Information Technology. | Digital infrastructure, international technology partnerships, and the industrial ecosystem around AI. | [^saudi-mcit] |
| Nigeria | **'Bosun Tijani** | Minister of Communications, Innovation and Digital Economy. | National AI strategy, local language/model capability, digital skills, and access to infrastructure. | [^nigeria] |
| South Africa | **Blade Nzimande** | Minister of Science, Technology and Innovation. | Public research funding, scientific capacity, and the connection between technology policy and industrial development. | [^south-africa] |
| Brazil | **Luciana Santos** | Minister of Science, Technology and Innovation. | The Brazilian AI plan, research infrastructure, domestic capability, and Portuguese-language AI. | [^brazil-mcti] |
| Brazil | **Waldemar Gonçalves Ortunho Júnior** | ANPD director-president. | Data-protection enforcement and its implications for AI training, biometric systems, and automated processing. | [^brazil-anpd] |

### International governance — distinguish the institutions

| Person | Public role or affiliation | Relevant authority or influence | Evidence |
| --- | --- | --- | --- |
| **Amandeep Singh Gill** | UN under-secretary-general and special envoy for digital and emerging technologies. | Global digital cooperation, institutional coordination, and the UN's AI-governance processes. | [^un-envoy] |
| **Yoshua Bengio** | Co-chair of the UN Independent International Scientific Panel on AI, in addition to his research roles above. | Evidence synthesis through the scientific panel; a different mandate from a national regulator or the International AI Safety Report. | [^un-panel] |
| **Maria Ressa** | Co-chair of the same UN scientific panel; journalist and Rappler cofounder. | Information integrity, democracy, and societal impacts in global evidence assessment. | [^un-panel] |
| **Doreen Bogdan-Martin** | ITU secretary-general; vice-chair of the AI for Good Global Commission. | International telecommunications coordination and the commission's work on trusted, inclusive AI adoption. | [^itu] |
| **Paul Kagame** | President of Rwanda; AI for Good Global Commission co-chair. | Political convening and African participation in the commission's international agenda. | [^itu] |
| **Marc Benioff** | Salesforce chair and CEO; AI for Good Global Commission co-chair. | Enterprise AI deployment and a public/private convening role; his corporate interests should be considered when assessing that role. | [^itu] |
| **Khaled El-Enany** | UNESCO director-general. | International work on AI ethics, education, science, and capacity building; older lists naming Audrey Azoulay are out of date. | [^unesco] |

The **UN scientific panel**, the **ITU-convened AI for Good Global Commission**, the **International AI Safety Report**, and the **EU AI Office** are separate initiatives. Similar names and overlapping participants do not give them the same membership, legal powers, or evidentiary functions.[^un-panel][^itu][^bengio][^sioli]

## How influence connects across these groups

Use this section as a research framework, not as a claim that any person secretly controls the whole system. Formal authority, economic dependence, scientific credibility, and public persuasion are different forms of influence.

| Decision or dependency | People to connect in this guide | Questions to investigate |
| --- | --- | --- |
| **What gets trained and released?** | Lab CEOs, chief scientists, compute leaders, and safety/behavior teams. | Who can change the research agenda? Who sets release criteria? Is a safety team's recommendation advisory, or can it stop a release? What does the public governance document actually require? |
| **Who can afford to build or run models?** | Huang, Su, Wei, Fouquet, Tan, Haas; cloud leaders; Son; Xiao; Amin. | Which constraints are chips, fabrication, networking, energy, financing, or contractual access? Do partnership announcements describe binding commitments, future plans, or available infrastructure? |
| **How do models reach users?** | Platform/product leaders such as Andreou, Kurian, Garman, Ternus, Federighi, and the Hugging Face founders. | Who controls defaults, distribution, APIs, permissions, interoperability, and switching costs? How do device-local and cloud deployment change privacy and access? |
| **What counts as progress or safety?** | Pachocki, Kaplan, Kavukcuoglu, Leike, Olah, Liang, Narayanan, METR researchers, Bengio, and government evaluators. | Are results reproducible? Do evaluations measure deployed workflows or curated benchmarks? Who funds the evaluator, controls model access, and chooses what is measured? |
| **Whose costs and benefits are visible?** | Gebru, Hanna, Miceli, Buolamwini, Mitchell, Chowdhury, Bender, Crawford, Whittaker, Kak, West, and regional researchers. | Which workers, languages, communities, and environmental costs are missing from aggregate performance or investment claims? Can affected people contest a decision or meaningfully opt out? |
| **Who can require a different outcome?** | Lawmakers, regulators, executive offices, procurement authorities, and courts—not just company policy teams or international advisers. | Which institution has jurisdiction? Is the relevant instrument a law, regulation, contract, voluntary commitment, scientific assessment, or political statement? What enforcement or appeal mechanism exists? |

Three distinctions are especially useful:

1. **Capability is not deployment authority.** Scientific leadership need not imply control of budgets, release decisions, or corporate governance. OpenAI's Foundation/Group structure is a concrete example of why these should be researched separately.[^openai-structure]
2. **“Open” is not a single property.** Ask separately about weights, training data, code, licenses, evaluation access, and platform governance. A model-sharing platform's promised openness after a proposed acquisition is not identical to an enforceable guarantee.[^hugging-face-deal]
3. **International participation is not international enforcement.** Scientific panels and convening commissions can influence evidence and agendas without acquiring the powers of domestic regulators.[^un-panel][^itu]

For a deeper profile, collect the person's official remit, actual decision rights, institution's funding/ownership, disclosed outside roles, recent technical or policy work, and credible critical assessments. Avoid inferring motives from affiliation alone.

## Maintenance and open verification questions

**Next scheduled review: 24 November 2026.** Refresh senior lab and government roles monthly and after major reorganizations, elections, resignations, or acquisitions. Reassess the wider research/public-interest roster at least quarterly. The date at the top records this document's research cutoff, not an automated live feed.

### Open verification questions

- **Announced researcher moves:** obtain a newer first-party start-date/remit confirmation for Shazeer and Jumper before removing the qualification in their rows.[^shazeer-move][^jumper-move]
- **Pending combinations:** require an explicit closing announcement for AMD/World Labs and NVIDIA/Hugging Face before assigning post-transaction titles or ownership. Do not let a passed expected closing date automatically change the roster.[^world-labs][^hugging-face-deal]
- **Meta's internal organization:** several functional roles rest on a 2025 transcript and reorganization reporting. Replace those with a current first-party functional directory or a newly dated, well-sourced account when available.[^meta-research]
- **Acting and newly created offices:** recheck CAISI's permanent leadership and the remit/membership of the October US AI task force. Coordination responsibilities should not be presented as newly enacted regulatory powers.[^caisi][^us-taskforce]
- **Conflicting official pages:** newer dated appointment notices can supersede a stale “about” page. Jade Leung's September 2026 transition is an example: the older AISI page still describes her former CTO/adviser roles.[^aisi][^leung-transition]
- **Mission leadership transitions:** preserve the distinction between Saurabh Vijay's current IndiaAI appointment and Abhishek Singh's earlier leadership. A source describing the February summit does not establish the October roster.[^india-policy][^indiaai-transition]
- **Geographic and occupational gaps:** expand coverage of Latin American and African research institutions, Southeast Asian developers, standards bodies, competition/data-protection authorities, labor organizations, and open-source maintainers. Present coverage is intentionally deeper for major model developers than for every application sector.

### Editing checklist

1. Verify the person's spelling, organization, actual title, effective date, and whether the position is permanent, interim, acting, advisory, or merely announced.
2. Prefer a current official directory plus a dated appointment/reorganization notice; use reputable reporting when the institution has not published the detail. Record unresolved contradictions instead of choosing the more impressive title.
3. Keep a concrete reason for inclusion. Specify a technical contribution, decision-making responsibility, public-interest role, or legal/policy remit—not simply fame or social-media following.
4. Separate current roles from historical achievements. Do not silently carry a title across a departure, company acquisition, ministry reorganization, or change in government.
5. Disclose consequential overlapping roles and distinguish similarly named institutions or people. Public-benefit status, nonprofit incorporation, and independence are different properties.
6. Add or update a source note with publisher, title/description, date when available, and URL. Check that every footnote reference resolves and remove unused definitions.
7. Update the snapshot date only after reviewing the whole roster; otherwise state exactly which entries received a partial update. Keep meaningful historical corrections in Git history rather than accumulating an unbounded alumni directory.
8. Check Markdown tables, section links, and `git diff --check` before merging. Publish only professional, publicly relevant information; this guide does not need private contact details or personal-life dossiers.

## Sources

Sources supporting the 24 October 2026 snapshot are listed below. **Primary** means an organizational directory, institutional biography, official document, or a person's own announcement; **reporting** means independent news coverage. A primary source establishes what an institution or person says; it does not independently validate capability claims. Live directories can change after this snapshot. Dated sources retain their own dates and should not be mistaken for a fresh employment confirmation.

[^openai-structure]: **Primary — OpenAI, [“Our structure”](https://openai.com/our-structure/).** Describes the Foundation/Group PBC relationship, current boards, and the Safety and Security Committee, including Kolter's nonvoting observer position on the Group board.

[^openai-reset]: **Reporting — Axios, [“OpenAI president Greg Brockman takes bigger role ahead of IPO”](https://www.axios.com/2026/08/14/openai-executive-greg-brockman-ipo), 14 August 2026; TIME, [Sam Altman interview and OpenAI's leadership reset](https://time.com/article/2026/08/26/openai-sam-altman-interview/), 26 August 2026.** Used for the expanded Brockman remit and departures; older operating rosters should not override these dated accounts.

[^openai-science]: **Reporting — MIT Technology Review, [“The two people shaping the future of OpenAI's research”](https://www.technologyreview.com/2025/07/31/1120885/the-two-people-shaping-the-future-of-openais-research/), 31 July 2025; TIME, [Altman interview](https://time.com/article/2026/08/26/openai-sam-altman-interview/), 26 August 2026.** Read with Mądry's October announcement for the more recent Pachocki affiliation.

[^madry-return]: **Primary — Aleksander Mądry, [announcement of his return to work with Jakub Pachocki](https://x.com/aleks_madry/status/2112231448103239998), 19 October 2026.** Establishes the announced safety/security work, not reinstatement as Preparedness head.

[^openai-research]: **Primary — OpenAI, [“Leadership updates”](https://openai.com/index/leadership-updates-march-2025/), 24 March 2025; [“The new ChatGPT Images is here”](https://openai.com/index/new-chatgpt-images-is-here/), 16 December 2025, contributor roles.** The former announces Chen's expanded chief research officer role; its Lightcap information is historical and is superseded by the 2026 reporting above.

[^openai-finance]: **Reporting — TIME, [Altman interview and OpenAI's corporate/financial leadership](https://time.com/article/2026/08/26/openai-sam-altman-interview/), 26 August 2026.** Identifies Friar as CFO; financing observations in the roster are an assessment of that remit, not a statement that announced capacity has been delivered.

[^openai-rajic]: **Primary — OpenAI, [“OpenAI appoints Dali Rajic as Chief Revenue Officer”](https://openai.com/index/dali-rajic-chief-revenue-officer/), 13 August 2026.** **Reporting — TechCrunch, [“OpenAI hires new CRO as executive shake-up continues”](https://techcrunch.com/2026/08/13/openai-hires-new-cro-as-executive-shake-up-continues/), 13 August 2026.**

[^weng-return]: **Reporting — TechCrunch, [Weng's departure from Thinking Machines and return to OpenAI](https://techcrunch.com/2026/07/29/thinking-machines-co-founder-lilian-weng-left-the-company-citing-health-reasons-then-joined-openai/), 29 July 2026.** Includes OpenAI's confirmation and description of her internal-research acceleration team; no personal health details are needed for this roster.

[^shazeer-move]: **Reporting — Reuters, [“Google's Gemini co-lead Noam Shazeer to join OpenAI”](https://www.reuters.com/technology/googles-gemini-co-lead-noam-shazeer-join-openai-2026-06-18/), 18 June 2026.** An announced move, not a source for a precise new executive title or confirmed start date.

[^anthropic-leadership]: **Primary — Anthropic, [leadership directory](https://www.anthropic.com/company/leadership).** Source for company executives, founders, and governance titles, including McCandlish's chief architect and Clark's public-benefit/Anthropic Institute remit.

[^anthropic-alignment]: **Primary — Anthropic, [Alignment Science research blog](https://alignment.anthropic.com/) and [“Teaching Claude Why”](https://alignment.anthropic.com/2026/teaching-claude-why/), 8 May 2026; Amanda Askell, [research biography](https://askell.io/).** The research and contributor list document alignment work; Askell describes her team's work on model character and finetuning. Organizational leadership and scientific contributions should not be conflated.

[^leike]: **Primary — Jan Leike, [research biography](https://jan.leike.name/).** Identifies his Alignment Science team leadership and describes scalable oversight, weak-to-strong generalization, and related research.

[^claude-code]: **Primary — Boris Cherny, [professional biography](https://borischerny.com/about/); Computer History Museum, [Cherny profile](https://computerhistory.org/profile/boris-cherny/).** These identify Cherny as Claude Code's creator (and the museum identifies him as its head); the inclusion assessment concerns the developer workflow, not an independently verified productivity claim.

[^karpathy-move]: **Reporting — TechCrunch, [“OpenAI co-founder Andrej Karpathy joins Anthropic's pre-training team”](https://techcrunch.com/2026/05/19/openai-co-founder-andrej-karpathy-joins-anthropics-pre-training-team/), 19 May 2026; Reuters, [appointment report](https://www.reuters.com/business/autos-transportation/former-tesla-ai-executive-openai-founding-member-andrej-karpathy-joins-anthropic-2026-05-19/), 19 May 2026.** Both include the pretraining-team context and company confirmation.

[^jumper-move]: **Reporting — Reuters, [“US scientist John Jumper to leave Google DeepMind for Anthropic”](https://www.reuters.com/technology/us-scientist-john-jumper-leave-google-deepmind-anthropic-2026-06-19/), 19 June 2026.** The announced break/move does not itself establish a later start date or title.

[^google-reorg]: **Primary — Sundar Pichai / Google, [“The next chapter of our AI momentum”](https://blog.google/company-news/inside-google/message-ceo/next-chapter-ai-momentum/), August 2026.** Source for the Kavukcuoglu/Hassabis reorganization and Alphabet/Google scientific-leadership changes.

[^discovery-loop]: **Primary — Jeff Dean, [Discovery Loop founding announcement](https://x.com/JeffDean/status/2085034604172603724), 5 August 2026.** **Reporting — GeekWire, [“The startup idea that convinced a UW computer science legend to leave Google after 27 years”](https://www.geekwire.com/2026/the-startup-idea-that-convinced-a-uw-computer-science-legend-to-leave-google-after-27-years/), 5 August 2026.** Identifies all four founders and the research-automation mission.

[^kohli]: **Primary — Google Research, [Pushmeet Kohli biography](https://research.google/people/105667/); Google, [author biography](https://blog.google/authors/pushmeet-kohli/).** These identify both the DeepMind and Cloud scientific roles.

[^gemini-safety]: **Primary — Anca Dragan and Raluca Ada Popa / Google, [“Developing Gemini 4 Argon responsibly and safely”](https://blog.google/innovation-and-ai/technology/safety-security/gemini-4-model-safety-security/), 13 October 2026.** The author bylines give their distinct safety/behavior and security titles; model-performance claims remain the company's claims.

[^meta-leadership]: **Primary — Meta, [leadership directory](https://www.meta.com/about/leadership/) and [Alexandr Wang biography](https://www.meta.com/about/leadership/alexandr-wang/).** Company-level titles and Wang's MSL leadership.

[^meta-research]: **Primary — Meta, [Q3 2025 follow-up call transcript](https://s21.q4cdn.com/399680738/files/doc_financials/2025/q3/META-Q3-2025-Follow-Up-Call-Transcript.pdf).** **Reporting — CNBC, [Zuckerberg's MSL creation memo](https://www.cnbc.com/2025/06/30/mark-zuckerberg-creating-meta-superintelligence-labs-read-the-memo.html), 30 June 2025.** These establish functional roles at the time; the roster explicitly flags the lack of a complete fresh 2026 internal directory.

[^ami]: **Reporting — TechCrunch, [“Yann LeCun's AMI Labs raises $1.03 billion to build world models”](https://techcrunch.com/2026/03/09/yann-lecuns-ami-labs-raises-1-03-billion-to-build-world-models/), 9 March 2026.** Source for AMI's founding leadership and technical roles, rather than LeCun's former Meta title.

[^microsoft-officers]: **Primary — Microsoft, [executive leadership biographies](https://news.microsoft.com/source/leadership/) and [“Introducing CoreAI – Platform and Tools”](https://blogs.microsoft.com/blog/2025/01/13/introducing-core-ai-platform-and-tools/), 13 January 2025.** The directory establishes Nadella and Smith's company-level titles; the dated announcement establishes Parikh's EVP/CoreAI remit.

[^microsoft-copilot]: **Primary — Microsoft, [“Announcing Copilot leadership update”](https://blogs.microsoft.com/blog/2026/03/17/announcing-copilot-leadership-update/), 17 March 2026.** Distinguishes Andreou's unified Copilot product leadership from Suleyman's frontier-model focus.

[^microsoft-ai-science]: **Reporting — GeekWire, [Microsoft's additional superintelligence-team hires](https://www.geekwire.com/2026/microsoft-adds-more-former-ai2-researchers-to-its-superintelligence-team/), 15 May 2026.** Identifies Simonyan's scientific leadership within Microsoft AI.

[^microsoft-ai2]: **Reporting — GeekWire, [Microsoft hires former Ai2 CEO Ali Farhadi and researchers](https://www.geekwire.com/2026/microsoft-hires-former-ai2-ceo-ali-farhadi-and-key-researchers-for-suleymans-ai-team/) and [additional hires](https://www.geekwire.com/2026/microsoft-adds-more-former-ai2-researchers-to-its-superintelligence-team/), 15 May 2026.** Covers Farhadi/Hajishirzi's move and continuing UW affiliations. Do not copy old Ai2 operating titles from academic biographies.

[^amazon-models]: **Primary — Amazon, [Andy Jassy's leadership update on Peter DeSantis and AI research](https://www.aboutamazon.com/news/company-news/andy-jassy-peter-desantis-amazon-leadership-update), 17 December 2025.** Includes Abbeel's leadership of the frontier-model team and DeSantis's expanded portfolio.

[^amazon-team]: **Primary — Amazon, [senior leadership / S-team directory](https://www.aboutamazon.com/news/workplace/amazon-s-team-members), updated August 2026.** Company-level roles for Jassy, DeSantis, and AWS CEO Garman.

[^amazon-lab-closure]: **Reporting — GeekWire, [“Amazon confirms it's closing key AI site in San Francisco but says work on its top models continues”](https://www.geekwire.com/2026/amazon-confirms-its-closing-key-ai-site-in-san-francisco-but-says-work-on-its-top-models-continues/), 24 July 2026.** Includes company confirmation of the site closure and continuation of frontier research under Abbeel; it does not establish the end of all Amazon model work.

[^xai]: **Reporting — Business Insider, [“xAI reorganizes its engineering team ahead of SpaceX IPO”](https://www.businessinsider.com/elon-musk-reorganizes-xai-ahead-of-spacex-ipo-2026-4), 9 April 2026; CNN, [SpaceX public-listing coverage](https://www.cnn.com/2026/06/12/business/live-news/spacex-goes-public-ipo), 12 June 2026.** Used for the SpaceX relationship, Musk, and Nicolls's president role; the guide deliberately uses a durable product/research label rather than assuming stable branding.

[^mistral]: **Primary — Mistral AI, [about and founders](https://mistral.ai/about/).** Explicitly lists Mensch as CEO, Lample as chief science officer, and Lacroix as CTO.

[^cohere]: **Primary — Cohere, [about and leadership](https://cohere.com/about).** Identifies the three cofounders and the current Pineau/Blunsom functional leadership.

[^deepseek]: **Primary — Chen Deli, [research biography](https://victorchen96.github.io/).** **Reporting — Reuters, [DeepSeek researcher's public appearance](https://www.reuters.com/world/asia-pacific/deepseek-researcher-pessimistic-over-ais-impact-startups-first-public-appearance-2025-11-07/), 7 November 2025, and [June financing report](https://www.reuters.com/world/asia-pacific/chinas-deepseek-closes-over-7-billion-funding-with-unusual-deal-structure-2026-06-16/), 16 June 2026.** Together document Chen's technical role and Liang's continuing CEO/founder role, without assigning Chen an unsupported chief-scientist title.

[^alibaba-april]: **Reporting — Caixin Global, [“Alibaba Revamps AI Structure With New Tech Committee”](https://www.caixinglobal.com/2026-04-08/alibaba-revamps-ai-structure-with-new-tech-committee-102431852.html), 8 April 2026; SCMP, [CEO-led technology committee](https://www.scmp.com/tech/article/3349428/alibaba-creates-ceo-led-technology-committee-amid-intensifying-ai-race), 8 April 2026.** For Wu Zeming, Li Feifei, and the April allocation of responsibilities; superseded for Zhou by June reporting.

[^alibaba-june]: **Reporting — SCMP, [“Alibaba forges Token Foundry as it hammers AI ambitions into shape”](https://www.scmp.com/tech/article/3356408/alibaba-forges-token-foundry-it-hammers-ai-ambitions-shape), 8 June 2026; 36Kr, [new division and research-institute structure](https://eu.36kr.com/en/p/3848411000902658), 11 June 2026.** Documents CEO Eddie Wu's consolidated AI effort and Zhou Jingren's chief-scientist/research-institute appointment. SCMP discloses that Alibaba owns the newspaper.

[^bytedance]: **Reporting — TechNode, [“Former Google DeepMind VP joins ByteDance as Seed team research lead”](https://technode.com/2025/02/24/former-google-deepmind-vp-joins-bytedance-as-seed-team-research-lead/), 24 February 2025; SCMP, [Wu Yonghui's foundational-research appointment](https://www.scmp.com/tech/big-tech/article/3299146/veteran-google-ai-researcher-joins-tiktok-owner-bytedance-lead-foundational-research), 18 February 2025.** Establishes the role and reporting relationship to Liang Rubo, not a formal chief-scientist title.

[^tencent-leadership]: **Primary — Tencent, [company information and management](https://www.tencent.com/en-us/about.html).** Source for Pony Ma and Martin Lau's corporate roles.

[^tencent-ai]: **Reporting — Bloomberg, [“Tencent Appoints Former OpenAI Researcher Its Chief AI Scientist”](https://www.bloomberg.com/news/articles/2025-12-17/tencent-appoints-former-openai-researcher-its-chief-ai-scientist), 17 December 2025; SCMP, [AI reorganization and Yao appointment](https://www.scmp.com/tech/big-tech/article/3336811/tencent-restructures-ai-operations-promotes-high-profile-recruit-chief-ai-scientist), 17 December 2025.** Documents the scientific role and model/infrastructure responsibilities.

[^baidu]: **Primary — Baidu Investor Relations, [Robin Li biography](https://ir.baidu.com/management/robin-li) and [Haifeng Wang biography](https://ir.baidu.com/management/haifeng-wang/).** Use the individual technical biography alongside the narrower main management directory.

[^xiaomi]: **Reporting — Reuters, [“Xiaomi to invest at least $8.7 billion in AI over next three years, CEO says”](https://www.reuters.com/world/asia-pacific/xiaomi-invest-least-87-billion-ai-over-next-three-years-ceo-says-2026-03-19/), 19 March 2026.** Identifies Lei Jun and Luo Fuli's leadership of MiMo; the spending figure is an announced plan, not used here as delivered capacity.

[^zhipu]: **Reporting — Bloomberg, [Zhipu's earnings and CEO Zhang Peng](https://www.bloomberg.com/news/articles/2026-03-31/zhipu-s-losses-climb-60-after-chinese-ai-rivalry-worsens), 31 March 2026; Forbes, [Zhipu chairman Liu Debing and chief scientist Tang Jie](https://www.forbes.com/sites/ywang/2026/06/02/zhipus-chairman-amasses-224-billion-fortune-as-chinese-ai-model-developer-soars-almost-1000/), 2 June 2026.** Used for roles, not the articles' wealth estimates.

[^moonshot]: **Reporting — Nikkei Asia, [“Moonshot AI's CEO turned down Apple and built Chinese rival to Anthropic”](https://asia.nikkei.com/business/technology/artificial-intelligence/moonshot-ai-s-ceo-turned-down-apple-and-built-chinese-rival-to-anthropic), 11 August 2026.** Identifies Yang's founder/CEO role and research background.

[^minimax]: **Primary — MiniMax Investor Relations, [management directory](https://ir.minimax.io/corporate-information/management); MiniMax / HKEX, [annual-meeting circular](https://www1.hkexnews.hk/listedco/listconews/sehk/2026/0519/2026051901411.pdf), 19 May 2026.** The dated filing identifies Yun as president and states that her COO tenure ended in March 2026, superseding the older live directory. The same sources establish Yan's combined roles and Zhao/Zhou's technical responsibilities.

[^stepfun]: **Reporting — Yicai Global, [StepFun financing and core management](https://www.yicaiglobal.com/news/chinese-ai-startup-stepfun-raises-usd7188-million-to-boost-foundation-models-development-terminal-agents-use), 26 January 2026; [subsequent management confirmation](https://www.yicaiglobal.com/news/stepfun-to-raise-nearly-usd25-billion-as-chinese-ai-startup-advances-hong-kong-ipo), 8 May 2026.** Names Jiang, Yin, Zhang, and Zhu with their distinct roles.

[^ssi]: **Primary — Safe Superintelligence, [company mission](https://ssi.inc/) and [updates](https://ssi.inc/updates), especially the 3 July 2025 leadership announcement.** Sutskever is formally CEO, Levy president, and Gross no longer part of SSI. The site also carries a July 2026 partnership update; the mission is not proof of achieved superintelligence.

[^thinking-machines]: **Primary — Thinking Machines Lab, [research/product mission](https://thinkingmachines.ai/); Mira Murati, [CTO transition announcement](https://x.com/miramurati/status/2011577319295692801), 14 January 2026.** **Reporting — TechCrunch, [Zoph/Metz departures and Chintala appointment](https://techcrunch.com/2026/01/14/mira-muratis-startup-thinking-machines-lab-is-losing-two-of-its-co-founders-to-openai/), 14 January 2026; Business Insider, [Weng's return to OpenAI and remaining Thinking Machines leadership](https://www.businessinsider.com/lilian-weng-returns-to-openai-after-leaving-thinking-machines-lab-2026-7), 29 July 2026.** The latter confirms Murati and Schulman's continuing CEO/chief scientist roles; these sources are not an unchanged original founding roster.

[^ineffable]: **Primary — Ineffable Intelligence, [mission and “A Note from Dave”](https://www.ineffable.ai/), note dated 15 January 2026.** Silver describes founding the organization and the reinforcement-learning agenda. Claims about a future superlearner are explicitly the company's ambitions.

[^recursive]: **Reporting — The New York Times, [“Notable Researchers Join $4 Billion Effort to Build Self-Improving A.I.”](https://www.nytimes.com/2026/05/13/technology/recursive-superintelligence-funding-ai.html), 13 May 2026; WIRED, [Recursive's research program and first model](https://www.wired.com/story/recursive-self-improvement-flint-latest-model-scary-good/), 21 October 2026.** Establishes the founding team and Socher's company leadership; claims of self-improvement require separate technical scrutiny.

[^periodic]: **Primary — Periodic Labs, [company mission and team](https://periodic.com/).** **Reporting/interview — Physics Today, [“Ekin Dogus Cubuk runs a startup to accelerate physics R&D using AI”](https://physicstoday.aip.org/news/ekin-dogus-cubuk-runs-a-startup-to-accelerate-physics-r-d-using-ai), 20 March 2026.** The latter explicitly identifies Çubuk as cofounder/co-CEO; the guide does not infer Fedus's precise current title from his earlier OpenAI position.

[^world-labs]: **Primary — World Labs, [AMD transaction announcement](https://www.worldlabs.ai/blog/amd-announcement), 28 September 2026.** The agreement and prospective Li title are conditional on closing. The announced plan retains Johnson/Mildenhall research leadership; it is not a completed-ownership announcement.

[^physical-intelligence]: **Primary — Karol Hausman, [research biography](https://karolhausman.github.io/); Sergey Levine, [Berkeley biography](https://people.eecs.berkeley.edu/~svlevine/); Chelsea Finn, [Stanford biography](https://ai.stanford.edu/~cbfinn/).** **Reporting — Bloomberg, [Physical Intelligence and CEO Hausman](https://www.bloomberg.com/news/articles/2025-11-20/robotics-startup-physical-intelligence-valued-at-5-6-billion-in-new-funding), 20 November 2025.** Used for affiliations and research focus, not an inference that commercial robot generality has been demonstrated.

[^ai2]: **Primary — Ai2, [Peter Clark biography](https://allenai.org/team/peterc) and [“Peter Clark Q&A”](https://allenai.org/blog/peter-clark-qa), 1 May 2026.** Identifies the interim CEO/scientific roles and Hilf's board chairmanship during the transition.

[^sakana]: **Primary — Sakana AI, [corporate information](https://sakana.ai/company-info/).** Current founder roles: Ha CEO, Jones CTO, Ito chairman. The separate older company page and early event biographies can still describe Ito as COO.

[^tii]: **Primary — TII, [Najwa Aaraj biography](https://www.tii.ae/team/dr-najwa-aaraj); [Falcon-Emirati announcement](https://www.tii.ae/insights/tii-launches-falcon-emirati-alongside-new-arabic-ai-models-speech-and-visual-text), 6 October 2026.** The latter confirms Aaraj and Al Bannai's distinct TII/ATRC responsibilities as of October.

[^sarvam]: **Primary — Sarvam AI, [about and founders](https://www.sarvam.ai/about-us).** Establishes both cofounders and their research/infrastructure backgrounds; it does not require choosing between inconsistent external CEO descriptions.

[^nvidia]: **Primary — NVIDIA, [Jensen Huang author biography](https://blogs.nvidia.com/blog/author/jen-hsun-huang/) and [executive biographies](https://nvidianews.nvidia.com/bios).** Establishes founder, president, and CEO roles.

[^amd]: **Primary — AMD, [leadership directory](https://www.amd.com/en/corporate/leadership.html).** Lisa Su's chair/CEO role; the separate World Labs announcement concerns a pending transaction.

[^tsmc]: **Primary — TSMC, [executives directory](https://www.tsmc.com/english/aboutTSMC/executives).** Wei's combined chairman/CEO role.

[^asml]: **Primary — ASML, [Board of Management](https://www.asml.com/en/company/governance/board-of-management).** Fouquet's president/CEO role and the distinction between executive management and supervisory governance.

[^broadcom]: **Primary — Broadcom Investor Relations, [Hock E. Tan biography](https://investors.broadcom.com/board-member/hock-e-tan).** President/CEO role and corporate background.

[^intel]: **Primary — Intel, [“Lip-Bu Tan Joins Intel as CEO”](https://newsroom.intel.com/corporate/tan-joins-intel-ceo), 18 March 2025.** Dated first-day/appointment account, not evidence that an older interim CEO remains in charge.

[^arm]: **Primary — Arm, [leadership](https://www.arm.com/company/leadership); SoftBank Group, [leadership announcement](https://group.softbank/en/news/press/20260421), 21 April 2026.** Read together for Haas's Arm and SoftBank Group International responsibilities.

[^huawei]: **Primary — Huawei, [Ren Zhengfei biography](https://www.huawei.com/en/executives/board-of-directors/ren-zhengfei).** Identifies founder/director/CEO status; do not confuse this with Huawei's rotating chairmanship.

[^google-cloud]: **Primary — Thomas Kurian / Google Cloud, [“Gemini at Work 2026: Introducing Gemini agent”](https://cloud.google.com/blog/products/ai-machine-learning/welcome-to-gemini-at-work-2026), 8 October 2026.** The byline establishes his CEO role; product and performance statements are the company's claims.

[^oracle]: **Primary — Oracle, [“Oracle Corporation Announces Promotion of Clay Magouyrk and Mike Sicilia”](https://www.oracle.com/news/announcement/oracle-corporation-announces-promotion-of-clay-magouyrk-and-mike-scilia-2025-09-22/), 22 September 2025.** Source for the co-CEO structure, Catz's executive vice-chair role, and Ellison's continuing chairman/CTO responsibilities.

[^apple]: **Primary — Apple, [leadership directory](https://www.apple.com/leadership/).** Current Ternus/Cook structure and the software, hardware, and legal/government-affairs roles. The CEO transition took effect on 1 September 2026; older Cook-as-CEO lists predate it.

[^coreweave]: **Primary — CoreWeave, [Mike Intrator biography](https://www.coreweave.com/leadership/mike-intrator).** Explicitly identifies cofounder, chairman, and CEO status; promotional claims about market coverage are not used as independent evidence.

[^softbank]: **Primary — SoftBank Group, [Masayoshi Son biography](https://group.softbank/en/about/officer/son), dated 24 June 2026.** Corporate chair/CEO role and relevant semiconductor/investment background.

[^g42]: **Primary — G42, [about and leadership](https://www.g42.ai/about).** Peng Xiao's group CEO role.

[^humain]: **Primary — HUMAIN, [CEO appointment announcement](https://x.com/HUMAINAI/status/1922176061988655281), May 2025, and [company overview](https://www.humain.com/).** The announcement establishes Amin's appointment; the overview describes the infrastructure, cloud, models, and applications strategy. Company plans are not independently verified operating capacity.

[^ibm]: **Primary — IBM, [Arvind Krishna biography](https://www.ibm.com/about/arvind).** Identifies chairman, president, and CEO and his AI/cloud/quantum remit. Darío Gil's current public-sector role is documented separately.

[^hugging-face]: **Primary — Thomas Wolf, [professional biography](https://thomwolf.io/); Julien Chaumond, [GitHub professional profile](https://github.com/julien-c).** **Reporting — CNBC, [Hugging Face/NVIDIA announcement and Delangue interview](https://www.cnbc.com/2026/09/03/nvidia-agrees-to-buy-hugging-face-for-almost-13-billion-ai-expansion.html), 3 September 2026.** These distinguish the three founders' CEO/CTO/CSO roles. Current company roles should not be rewritten as post-acquisition appointments before closing.

[^linux]: **Primary — Linux Foundation, [leadership and fellows](https://www.linuxfoundation.org/about/leadership).** Torvalds's fellowship and Linux-creator role; kernel development is collaborative, not the work of one person alone.

[^hugging-face-deal]: **Primary — NVIDIA, [Hugging Face acquisition-agreement announcement](https://blogs.nvidia.com/blog/nvidia-to-acquire-hugging-face/), 3 September 2026; [Form 8-K](https://www.sec.gov/Archives/edgar/data/1045810/000104581026000078/nvda-20260902.htm), event dated 2 September 2026.** **Reporting — CNN, [acquisition-agreement coverage](https://www.cnn.com/2026/09/03/tech/nvidia-hugging-face-ai-acquisition), 3 September 2026.** The filing specifies expected first-half 2027 closing and regulatory conditions. “Agreed to acquire” is not “has acquired.”

[^bengio]: **Primary — LawZero, [Yoshua Bengio biography](https://lawzero.org/en/team/yoshua-bengio).** Identifies LawZero co-president/scientific director, Université de Montréal professor, Mila founder/scientific adviser, and International AI Safety Report chair. These roles are separate from the UN panel below.

[^hinton]: **Primary — University of Toronto, [Geoffrey Hinton biography](https://www.cs.toronto.edu/~hinton/).** Current emeritus affiliation and selected research; former Google employment is historical.

[^sutton]: **Primary — Amii, [Richard S. Sutton biography](https://www.amii.ca/people/richard-s-sutton).** **Reporting — BetaKit, [“AI pioneer Richard Sutton founds new research lab”](https://betakit.com/ai-pioneer-richard-sutton-founds-new-research-lab/), 20 July 2026.** The latter supplies the Oak Lab/Keen transition; older Keen biographies do not establish current employment there.

[^russell]: **Primary — UC Berkeley, [Stuart Russell faculty biography](https://www2.eecs.berkeley.edu/Faculty/Homepages/russell.html).** Institutional role, CHAI leadership, and research areas.

[^liang]: **Primary — Stanford, [Percy Liang profile](https://profiles.stanford.edu/percy-liang); [professoriate appointments report](https://news.stanford.edu/stories/2026/04/report-of-the-president-academic-council-professoriate-appointments), April 2026.** The latter records his promotion to full professor effective 1 April 2026; older associate-professor descriptions are stale.

[^choi]: **Primary — Stanford, [Yejin Choi profile](https://profiles.stanford.edu/yejin-choi).** Faculty/HAI roles and the recorded partial leave beginning 1 October 2026. A leave notation is not a departure from the university.

[^ng]: **Primary — Andrew Ng, [professional biography](https://www.andrewng.org/).** DeepLearning.AI, AI Fund, education, and Stanford roles.

[^song]: **Primary — UC Berkeley, [Dawn Song faculty biography](https://www2.eecs.berkeley.edu/Faculty/Homepages/song.html); [Responsible Decentralized Intelligence center](https://rdi.berkeley.edu/).** Institutional affiliation and security/privacy/decentralization research.

[^marivate]: **Primary — University of Pretoria, [Marivate's UN scientific-panel appointment and institutional biography](https://www.up.ac.za/news/ups-prof-vukosi-marivate-appointed-un-independent-international-scientific-panel-ai), 11 March 2026; UN, [scientific-panel member biographies](https://www.un.org/independent-international-scientific-panel-ai/en/panel-members).** Documents his university/center roles and Lelapa/Masakhane research-community affiliations.

[^lelapa]: **Primary — Lelapa AI, [current team directory](https://lelapa.ai/team); Wits University, [Rosman's AI-policy panel appointment](https://www.wits.ac.za/news/latest-news/research-news/2026/2026-05/professor-benjamin-rosman-leads-south-africas-ai-policy-panel.html), 14 May 2026, and [Rosman faculty/research profile](https://www.wits.ac.za/news/latest-news/research-news/2025/2025-07/mind-director-named-coveted-cifar-fellow.html), 21 July 2025.** The current company directory lists Abbott as CTO, not her older COO title; Wits documents Rosman's academic, MIND, and Indaba roles.

[^metr]: **Primary — METR, [about, leadership, funding, and independence disclosures](https://metr.org/about).** Names Barnes, Painter, Wijk, and Rush; separately explains funding, model access, free tokens, and conflicts-of-interest safeguards. Read the disclosures rather than assuming independence solely from incorporation status.

[^hendrycks]: **Primary — Center for AI Safety, [about and team](https://safe.ai/about); Dan Hendrycks, [professional biography and Meta update](https://danhendrycks.com/).** The institutional site supplies the director title; his own site supplies the additional Meta safety/evaluation work and advisory relationships.

[^narayanan]: **Primary — Princeton, [Arvind Narayanan faculty biography](https://www.cs.princeton.edu/~arvindn/).** Computer-science/CITP roles and research on privacy, security, and evaluation.

[^dair]: **Primary — Distributed AI Research Institute, [team](https://dair-institute.org/team/).** Identifies Gebru as executive director, Hanna as research director, and Miceli as research lead; includes their research focus and public biographies.

[^buolamwini]: **Primary — Algorithmic Justice League, [about](https://www.ajl.org/about).** Founder and research/advocacy context for Buolamwini's work on algorithmic discrimination and biometric systems.

[^mitchell]: **Primary — Margaret Mitchell, [professional biography](https://m-mitchell.com/).** Hugging Face chief ethics scientist role and research on evaluation/documentation.

[^chowdhury]: **Primary — Humane Intelligence PBC, [company and leadership](https://humane-intelligence.com/); Humane Intelligence nonprofit, [history and leadership transition](https://humane-intelligence.org/about/).** Establishes her PBC founder/CEO role and separate nonprofit cofounder status. The nonprofit now has its own executive director; do not merge these organizations into one biography.

[^bender]: **Primary — University of Washington, [Emily M. Bender faculty biography](https://faculty.washington.edu/ebender/).** Current linguistic research/laboratory roles and selected work.

[^crawford]: **Primary — Kate Crawford, [professional biography and projects](https://katecrawford.net/).** USC and Knowing Machines affiliations; research on AI's political, material, and social foundations.

[^whittaker]: **Primary — Signal Foundation, [current board and leadership](https://signalfoundation.org/); Signal, [“A Message from Signal's New President”](https://signal.org/blog/announcing-signal-president/), 6 September 2022; AI Now Institute, [about and founding history](https://ainowinstitute.org/about).** The directory supplies a current president affiliation, the dated announcement establishes the appointment, and the institute documents the separate cofounder affiliation.

[^ai-now]: **Primary — AI Now Institute, [about and leadership](https://ainowinstitute.org/about).** Kak/Myers West co-executive leadership and the institute's public-interest research program.

[^us-executive]: **Primary — White House, [administration directory](https://www.whitehouse.gov/administration/).** President Trump and Vice President Vance's current offices; political descriptions on an administration website are not neutral independent analysis.

[^us-pcast]: **Primary — White House, [PCAST appointments](https://www.whitehouse.gov/releases/2026/03/president-trump-announces-appointments-to-presidents-council-of-advisors-on-science-and-technology/), March 2026.** For the external advisory council and Sacks/Kratsios co-chair roles, not a claim that PCAST is a regulator.

[^us-taskforce]: **Reporting — Semafor, [“White House's AI task force gets to work”](https://www.semafor.com/article/10/15/2026/white-houses-ai-task-force-gets-to-work), 15 October 2026.** Reports Clayton's DNI/task-force chair role, the Michael/Ferguson/Kupor vice-chairs, Cairncross's work, membership, and the adopted charter. The article expressly notes uncertainty about enforcement and current-law application.

[^ostp]: **Primary — White House OSTP, [“The Director”](https://www.whitehouse.gov/ostp/the-director/).** Kratsios's standing OSTP and presidential science/technology-adviser responsibilities, distinct from his PCAST co-chairmanship.

[^sacks-transition]: **Reporting — CNBC, [“David Sacks says his time as Trump's crypto and AI czar has ended”](https://www.cnbc.com/2026/03/26/david-sacks-trump-crypto-ai-czar.html), 26 March 2026; Reuters, [transition to advisory role](https://www.reuters.com/world/us/white-house-ai-czar-sacks-step-down-moves-advisory-role-2026-03-27/), 27 March 2026.** His limited special-government-employee appointment should not be treated as a continuing full-time White House office.

[^ftc]: **Primary — FTC, [Andrew N. Ferguson biography](https://www.ftc.gov/about-ftc/commissioners-staff/andrew-n-ferguson).** FTC chairmanship; task-force duties are sourced separately above.

[^commerce]: **Primary — White House, [Cabinet](https://www.whitehouse.gov/administration/the-cabinet/); BIS, [leadership and offices](https://www.bis.gov/about-bis/bis-leadership-and-offices); Department of Commerce, [BIS overview](https://www.commerce.gov/bureaus-and-offices/bis).** Sources for Lutnick's cabinet role and Kessler's distinct export-control portfolio. The broad Commerce leadership page was temporarily under maintenance during the final source check, so the more specific directories are supplied here.

[^nist]: **Primary — NIST, [“Arvind Raman Confirmed as the 18th NIST Director”](https://www.nist.gov/news-events/news/2026/07/arvind-raman-confirmed-18th-nist-director), July 2026; [Office of the Director](https://www.nist.gov/director).** Confirms the 30 June swearing-in and standing standards/measurement role, rather than perpetuating a predecessor's acting title.

[^caisi]: **Reporting — Axios, [AI security agency leadership departure](https://www.axios.com/2026/07/20/trump-ai-security-agency-head-resigns), 20 July 2026.** Establishes Fall's departure and Raman's acting CAISI assignment, distinct from Raman's permanent NIST directorship.

[^fcc]: **Primary — FCC, [Brendan Carr and current statements](https://www.fcc.gov/about/leadership/brendan-carr/statements?year=2026).** **Reporting — FedScoop, [Carr on the executive-order assignment concerning state AI laws](https://fedscoop.com/fcc-chair-brendan-carr-trump-order-banning-state-ai-rules/).** A proceeding/assignment is not proof that a final preemptive rule has been adopted or upheld.

[^doe]: **Primary — Department of Energy, [Darío Gil biography](https://www.energy.gov/person/dario-gil) and [Genesis Mission](https://www.energy.gov/undersecretaryforscience/genesis-mission/genesis-mission).** Current public office and DOE mission-director role; do not list him using his former IBM Research title.

[^senate-commerce]: **Primary — US Senate Commerce Committee, [about](https://www.commerce.senate.gov/about/) and [members](https://www.commerce.senate.gov/about/members/).** Current Cruz/Cantwell leadership and committee jurisdiction.

[^house-ai]: **Primary — Representative Obernolte, [House AI Task Force final report](https://obernolte.house.gov/AITFReport), December 2024.** Documents the Obernolte/Lieu-led work and recommendations. The guide does not convert that completed task force into a permanent institution or describe its recommendations as enacted law.

[^california]: **Primary — California governor, [SB 53 signing announcement](https://www.gov.ca.gov/2025/09/29/governor-newsom-signs-sb-53-advancing-californias-world-leading-artificial-intelligence-industry/), 29 September 2025; California Legislature, [SB 53 text and status](https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=202520260SB53).** Names Newsom and author Wiener and distinguishes enacted SB 53 from earlier, vetoed legislation.

[^uk-pm]: **Primary — GOV.UK, [Prime Minister: current officeholder, responsibilities, and biography](https://www.gov.uk/government/ministers/prime-minister).** Explicitly identifies Burnham and his 20 July 2026 start date; this supersedes an older Starmer-era roster.

[^uk-ai-minister]: **Primary — GOV.UK, [Kanishka Narayan biography and responsibilities](https://www.gov.uk/government/people/kanishka-narayan).** Explicitly records the 20 July 2026 joint Cabinet Office/BIST appointment and the AI Strategy, AI Taskforce, AISI, infrastructure, science, and adoption portfolios.

[^aisi]: **Primary — AI Security Institute, [about and leadership](https://www.aisi.gov.uk/about).** Hogarth/Summerfield roles and the institute's research remit. Its older Leung CTO description is superseded by the dated September notice below.

[^leung-transition]: **Primary — GOV.UK, [“Dr Jade Leung appointed Vice-Chair of the AI Security Institute and Security Advisor to the AI Taskforce”](https://www.gov.uk/government/news/dr-jade-leung-appointed-vice-chair-of-the-ai-security-institute-and-security-advisor-to-the-ai-taskforce), 23 September 2026.** Specifies stepping back from the full-time CTO/prime-ministerial-adviser roles at the end of September and continuing strategic security work.

[^eu-college]: **Primary — European Commission, [College of Commissioners](https://commission.europa.eu/about/organisation/college-commissioners_en).** Commission presidency and division of political portfolios; the Commission is not the entirety of EU legislative government.

[^virkkunen]: **Primary — European Commission, [Henna Virkkunen biography and portfolio](https://commission.europa.eu/about/organisation/college-commissioners/henna-virkkunen_en).** Tech sovereignty, security, democracy, and digital-policy responsibilities.

[^sioli]: **Primary — European Innovation Council, [Lucilla Sioli biography](https://eic.ec.europa.eu/lucilla-sioli_en); European Commission, [Apply AI Summit speakers](https://digital-strategy.ec.europa.eu/en/miscellaneous/meet-speakers-apply-ai-summit), August 2026; [European AI Office](https://digital-strategy.ec.europa.eu/en/policies/ai-office).** Read together for the director role and office mandate.

[^dg-connect]: **Primary — European Commission, [DG CONNECT department, leadership, and organization chart](https://commission.europa.eu/about/departments-and-executive-agencies/communications-networks-content-and-technology_en), including the chart dated 1 October 2026.** Confirms Viola rather than assuming a rumored successor has taken office.

[^ribera]: **Primary — European Commission, [Teresa Ribera biography and portfolio](https://commission.europa.eu/about/organisation/college-commissioners/teresa-ribera_en).** Competition responsibilities are part of the Clean, Just and Competitive Transition portfolio, separate from the AI Office.

[^france]: **Primary — Élysée, [Emmanuel Macron](https://www.elysee.fr/en/emmanuel-macron).** Presidential office and public policy/diplomatic work; inclusion concerns national political direction, not personal control of a private lab.

[^germany]: **Primary — German Federal Ministry for Digital Transformation and Government Modernisation, [Wildberger at ALL IN 2026 in Canada](https://bmds.bund.de/aktuelles/aktuelle-meldungen/detail/bundesminister-wildberger-auf-der-all-in-2026-in-kanada), September 2026.** Current ministerial role and AI/digital-sovereignty engagement.

[^canada-strategy]: **Primary — Prime Minister of Canada, [“Prime Minister Carney launches new National Council on Artificial Intelligence”](https://www.pm.gc.ca/en/news/news-releases/2026/10/02/prime-minister-carney-launches-new-national-council-artificial), 2 October 2026.** Current prime minister, national strategy, and advisory-coordination structure; economic and jobs targets are policy ambitions, not realized outcomes.

[^canada-minister]: **Primary — Innovation, Science and Economic Development Canada, [Solomon's UK/Ireland visit and AI cooperation](https://www.canada.ca/en/innovation-science-economic-development/news/2026/10/minister-solomon-concludes-successful-visit-to-the-united-kingdom-and-ireland-to-advance-canadas-global-ai-leadership.html), 15 October 2026.** Confirms his current title and the international security/adoption portfolio.

[^china-state]: **Primary — State Council / Xinhua, [Xi Jinping's WAIC keynote](https://english.www.gov.cn/news/202607/17/content_WS6a59bb10c6d00ca5f9a0c43d.html), 17 July 2026; [Li Qiang's State Council executive meeting](https://english.www.gov.cn/news/202607/31/content_WS6a6ca61cc6d00ca5f9a0c83a.html), 31 July 2026; [current State Council portal](https://english.www.gov.cn/).** Establishes the political leadership and policy agenda; official claims about national capabilities are not independent assessments.

[^china-ding]: **Primary — State Council / Xinhua, [Ding on AI self-reliance and open cooperation](https://english.www.gov.cn/news/202604/29/content_WS69f202bcc6d00ca5f9a0ab41.html), 29 April 2026; [data-enterprise inauguration](https://english.www.gov.cn/news/202610/20/content_WS6ad76d74c6d00ca5f9a0dda0.html), 20 October 2026.** Confirms vice-premier/Politburo Standing Committee roles and continuing data/AI policy involvement.

[^china-cac]: **Primary — CAC, MIIT, and Ministry of Public Security, [Measures for Network Data Security Risk Assessment (网络数据安全风险评估办法), Order No. 24](https://www.cac.gov.cn/2026-06/18/c_1783525609815499.htm), 18 June 2026, effective 20 August 2026 (Chinese).** The official signatures identify Zhuang as CAC director and Li as MIIT minister and illustrate the separate agencies' data-governance responsibilities.

[^india-policy]: **Primary — Press Information Bureau / MeitY, [India AI Impact Summit compute and policy announcement](https://www.pib.gov.in/PressReleseDetailm.aspx?PRID=2229171&reg=3&lang=2), 17 February 2026.** Documents Modi's strategic framing and Vaishnaw/Krishnan's roles. Its Singh-as-IndiaAI-CEO description was accurate for that event but is superseded by the June appointment below.

[^india-october]: **Reporting — Economic Times, [“Govt to issue consultation paper on AI law in coming weeks: Vaishnaw”](https://economictimes.indiatimes.com/tech/technology/govt-to-issue-consultation-paper-on-ai-law-in-coming-weeks-vaishnaw/articleshow/135152578.cms), 23 October 2026.** Establishes a proposed consultation, not a completed AI statute.

[^krishnan-current]: **Reporting — Economic Times, [“Centre recognises AI needs more regulation: IT Secretary”](https://economictimes.indiatimes.com/ai/ai-insights/centre-recognises-ai-needs-more-regulation-it-secretary/articleshow/135154257.cms), 23 October 2026.** Supplies a fresh MeitY-secretary affiliation; policy plans remain proposals.

[^indiaai-transition]: **Reporting — ETGovernment, [Saurabh Vijay given additional charge as IndiaAI Mission CEO](https://government.economictimes.indiatimes.com/news/digital-india/uidai-chief-saurabh-vijay-given-additional-charge-as-ceo-of-indiaai-mission/131464210), 2 June 2026; ANI, [Vijay takes charge](https://www.aninews.in/news/business/uidai-ceo-saurabh-vijay-takes-charge-as-ceo-of-indiaai-mission20260605231335/), 5 June 2026.** The latter quotes the mission's announcement. The additional charge is distinct from his UIDAI role and supersedes Abhishek Singh's earlier mission leadership.

[^japan-cabinet]: **Primary — Prime Minister's Office of Japan, [Takaichi's cabinet-reshuffle press conference](https://japan.kantei.go.jp/105/speech/202609/0917kaiken.html), 17 September 2026.** The full transcript distinguishes Furukawa's digital/AI strategy portfolios from Akazawa's METI and AI social-implementation responsibilities.

[^korea]: **Primary — Republic of Korea MSIT, [“MSIT Begins ‘AI for All Project’ to Offer AI Services through a Sovereign AI Model”](https://www.msit.go.kr/eng/bbs/view.do?bbsSeqNo=42&mId=4&mPid=2&nttSeqNo=1285&sCode=eng), 2026.** Identifies Bae as deputy prime minister and minister and gives a concrete domestic-model/adoption program.

[^singapore]: **Primary — Singapore PMO, [Cabinet appointments](https://www.pmo.gov.sg/about-us/the-cabinet/), updated 27 July 2026; Cyber Security Agency, [Josephine Teo's ASEAN cybersecurity remarks](https://www.csa.gov.sg/news-events/speeches/opening-remarks-by-minister-josephine-teo-at-the-11th-asean-ministerial-conference-on-cybersecurity-on-14-october-2026/), 14 October 2026.** Current ministerial office and regional AI/cybersecurity engagement.

[^uae]: **Primary — UAE Cabinet, [“Mohammed bin Rashid approves establishing Artificial Intelligence and Data Authority”](https://uaecabinet.ae/en/news/mohammed-bin-rashid-approves-establishing-artificial-intelligence-and-data-authority).** Identifies Al Olama's authority chairmanship and ministerial/Prime Minister's Office responsibilities, distinct from emirate-level institutions.

[^sdaia]: **Primary — Saudi Data and AI Authority, [Year of Artificial Intelligence](https://sdaia.gov.sa/en/MediaCenter/Pages/ai-year.aspx), 2026.** Confirms Alghamdi as SDAIA president and describes the national coordination agenda.

[^saudi-mcit]: **Primary — Saudi Ministry of Communications and Information Technology, [Abdullah Alswaha biography](https://mcit.gov.sa/en/he-eng-abdullah-alswaha-minister-communications-information-technology).** Ministerial role and digital-infrastructure remit.

[^nigeria]: **Primary — Nigerian Federal Ministry of Communications, Innovation and Digital Economy, [Tijani named a 2026 Goalkeepers Champion](https://fmcide.gov.ng/honourable-minister-dr-bosun-tijani-named-2026-gates-foundation-goalkeepers-champion/), 2026.** A current institutional affiliation and account of his digital/AI policy work; the award itself is not the criterion for inclusion.

[^south-africa]: **Primary — South African Government, [Minister Blade Nzimande at the FuturePharma facility launch](https://www.gov.za/news/speeches/minister-blade-nzimande-launch-futurepharma-facility-07-oct-2026), 7 October 2026.** Confirms the current science/technology/innovation portfolio rather than relying on older cabinet lists.

[^brazil-mcti]: **Primary — Brazilian MCTI, [“Quem é quem” institutional directory](https://www.gov.br/mcti/pt-br/acesso-a-informacao/institucional/quem-e-quem), updated 28 May 2026 (Portuguese).** Identifies Luciana Santos as minister. The roster concerns her office as of the cutoff, not assumptions about subsequent electoral or cabinet outcomes.

[^brazil-anpd]: **Primary — ANPD, [Waldemar Gonçalves Ortunho Júnior biography and institutional directory](https://www.gov.br/anpd/pt-br/acesso-a-informacao/institucional/quem-e-quem/diretor-presidente-waldemar-goncalves-ortunho-junior) (Portuguese).** Source for the director-president affiliation and professional background.

[^un-envoy]: **Primary — UN Office for Digital and Emerging Technologies, [leadership](https://www.un.org/digital-emerging-technologies/leadership).** Gill's under-secretary-general/special-envoy mandate dates from 1 January 2025; it is broader than his former Envoy on Technology title.

[^un-panel]: **Primary — United Nations, [Independent International Scientific Panel on AI FAQ](https://www.un.org/independent-international-scientific-panel-ai/en/faq) and [panel members](https://www.un.org/independent-international-scientific-panel-ai/en/panel-members).** Membership, Bengio/Ressa co-chairmanship, and the scientific-assessment mandate; this is not the ITU commission or the International AI Safety Report.

[^itu]: **Primary — ITU, [AI for Good Global Commission announcement](https://www.itu.int/en/mediacentre/Pages/PR-2026-07-02-AI-for-Good-Global-Commission.aspx), 2 July 2026.** Names Kagame and Benioff as co-chairs and Bogdan-Martin as vice-chair and sets out the convening mission. Corporate participation is not evidence of regulatory authority.

[^unesco]: **Primary — UNESCO, [Director-General Khaled El-Enany](https://www.unesco.org/en/director-general).** Records his assumption of office on 15 November 2025 and current institutional responsibilities; the former Azoulay title should not be carried forward.