Jack Clark

wiki
, , , ,
1

Co-founder and Head of Policy of Anthropic, former Policy Director of OpenAI, and author of Import AI — the weekly newsletter that has served as the field’s primary research intelligence digest since 2016, read by over 116,000 subscribers — who occupies a singular position at the intersection of frontier AI development, safety advocacy, and international governance.

Profile

Field Detail
Full name Jack Clark
Date of birth Not publicly available
Place of birth Brighton, England
Nationality British
Current institution Anthropic
Current role Co-founder & Head of Policy
Newsletter Import AI (importai.substack.com / jack-clark.net)
Personal website jack-clark.net
X / Twitter @jackclarkSF
GitHub jackclarksf
Google Scholar wjkaNgcAAAAJ

Overview

Jack Clark is co-founder and Head of Policy of Anthropic, an AI safety company he helped establish in 2021 alongside Dario Amodei, Daniela Amodei, and others who had departed OpenAI. He previously served as Policy Director at OpenAI (2016–2020), entering the AI industry after a career as a technical journalist who, by his own account, was the world’s only dedicated distributed systems reporter at The Register and Bloomberg’s sole neural network correspondent. Since 2016 he has written Import AI, a weekly newsletter synthesizing AI research for a specialist audience that has grown from a small mailing list to over 116,000 subscribers by early 2026. Clark was a founding member of the Stanford AI Index (2017–2024), an inaugural member of the US National Artificial Intelligence Advisory Committee (NAIAC, 2021–2024), and in July 2023 briefed the UN Security Council at its first formal meeting on AI threats to international peace and security. Named one of TIME’s 100 Most Influential People in AI in 2023, he has become one of the most prominent advocates for democratic oversight of frontier AI development, arguing consistently that governments must hold AI companies accountable through robust evaluation frameworks.

Early Life & Education

Clark was born in Brighton, England, and attended Varndean College in Brighton for his secondary education. He completed a B.A. in English Literature with Creative Writing at the University of East Anglia (2006–2009). This humanities background — unusual for the senior leadership of a frontier AI company — has shaped both his communication style and his orientation toward making technical systems legible to non-specialist audiences, a theme he has articulated explicitly as the animating purpose of his career. After graduating, he began work as a technology copywriter at Adfero in London, before transitioning into technology journalism.

Career

Adfero — Researcher and Reporter (September 2009–June 2010)

Technology copywriting and research at a London content and media agency; Clark’s entry point into covering technology.

ZDNet UK / CBS Interactive (UK) — Reporter (July 2010–January 2013)

Covered UK enterprise technology from London, developing early expertise in distributed systems, data centers, cloud infrastructure, and quantum computing.

The Register — Reporter (February 2013–August 2014)

Relocated to San Francisco. At The Register — a long-running UK-founded enterprise IT publication — Clark became what has been described as the world’s only dedicated distributed systems reporter, covering cloud infrastructure, networking, and early AI research with an unusually technical eye unusual among general-audience tech journalists.

Bloomberg LP — Reporter (August 2014–August 2016)

Joined Bloomberg in San Francisco initially to cover enterprise technology companies including Oracle, Dell, Hewlett-Packard, and Salesforce. His beat evolved rapidly: he became Bloomberg’s dedicated neural network and AI reporter, conducting some of the earliest major-press interviews with researchers including Dario Amodei on the then-nascent AI safety agenda. This role placed him at the center of the emerging deep learning industry at the precise moment AlphaGo, GPT precursors, and LSTM-based systems were beginning to attract mainstream attention.

OpenAI — Strategy & Communications Director → Policy Director (September 2016–2020)

Left Bloomberg in September 2016 to join OpenAI, then a newly incorporated non-profit, as Strategy and Communications Director — a role that encompassed community outreach, policy, communications, and strategic positioning. He later became Policy Director, the most senior policy role at the organization. In 2018 he testified before the US House Committee on Oversight and Government Reform on AI policy. Also in 2018, he co-authored one of the first systematic analyses of AI misuse risks, “The Malicious Use of Artificial Intelligence” (arXiv:1802.07228), with researchers from Cambridge, OpenAI, Oxford, and the EFF. During his OpenAI tenure he was also a founding member of the Stanford AI Index (2017), a data-driven initiative tracking global AI progress.

Import AI launched during this period in 2016 as a personal newsletter, growing organically from a small specialist mailing list into a major weekly publication while Clark was simultaneously shaping OpenAI’s policy posture.

Anthropic — Co-founder & Head of Policy (2021–present)

Clark was among the founding cohort of Anthropic, which incorporated in 2021. He serves as Head of Policy, responsible for shaping Anthropic’s external policy positions, regulatory engagement, and public communications on AI safety and governance. His policy stance is notably skeptical of purely industry-driven governance: in his July 2023 address to the UN Security Council — the body’s first formal meeting on AI and international security — he stated, “We cannot leave the development of artificial intelligence solely to private-sector actors,” and called for governments to develop mandatory evaluation systems for AI capabilities, misuses, and safety flaws.

In March 2026, Clark shifted the focus of his Anthropic role to concentrate primarily on public communication of AI risk — an evolution that coincides with Import AI’s rapid subscriber growth and his increasing presence at international forums including the 2026 Bilderberg conference.

Key Contributions

  • Import AI newsletter (2016–present) — A weekly research digest tracking arXiv preprints, policy developments, and industry milestones, read by 116,000+ subscribers as of early 2026 (crossing 100,000 in January 2026). Each issue combines technical summaries with a brief embedded science fiction story in which Clark explores speculative AI futures. Import AI is widely regarded as the field’s most influential specialist newsletter and a primary channel through which researchers, policymakers, and investors track AI progress. Clark has described running it as a mechanism for solving “information asymmetries” between technical insiders and the broader public.

  • “The Malicious Use of Artificial Intelligence” (Brundage, Avin, Clark et al.; arXiv:1802.07228, 2018) — Co-authored with a cross-institutional group from Cambridge, Oxford, OpenAI, and the EFF; one of the first systematic attempts to map how AI systems could be weaponized by malicious actors across digital, physical, and political domains. The paper remains a foundational reference in AI safety and dual-use risk research.

  • Stanford AI Index (founding member, 2017–2024) — Co-established the annual report tracking quantitative and qualitative measures of AI progress, widely used by governments, journalists, and researchers to ground policy discussions in empirical data rather than hype or fear.

  • UN Security Council briefing (July 18, 2023) — One of two non-governmental experts (alongside Yi Zeng of the Chinese Academy of Sciences) to brief the Council’s first formal session on AI and international peace, the other briefer being UN Secretary-General António Guterres. Clark used the platform to call for international investment in AI evaluation infrastructure and democratic accountability for AI companies.

  • US National Artificial Intelligence Advisory Committee (NAIAC) (inaugural member, 2021–2024) — Served as one of the inaugural members of the US government’s primary AI advisory body, advising the President and the National AI Initiative Office on issues including AI safety, governance frameworks, and the use of AI in federal systems.

  • OECD ONE AI — Co-chaired the OECD’s working group on classification and definition of AI systems, contributing to the international definitional and governance frameworks that underpin regulatory discussions across member states.

  • Anthropic’s policy positioning — Under Clark’s policy leadership, Anthropic has taken unusually explicit public positions on AI governance risks — including advocating for international AI testing regimes, supporting the Bletchley Declaration (2023 AI Safety Summit), and publishing research on model capabilities evaluation — establishing the company as distinct from competitors in its willingness to engage with existential risk arguments.

Awards & Recognition

  • TIME 100 Most Influential People in AI (2023) — Listed among TIME magazine’s inaugural AI 100 for his combined roles as Anthropic co-founder, policy advocate, and author of Import AI
  • Bilderberg Conference delegate (2026) — Among a small cohort of AI figures invited to the annual private conference of political, business, and intellectual leaders

Key Relationships

  • Dario Amodei — CEO of Anthropic and co-founder; the two overlapped at Bloomberg (Clark interviewed Amodei while at Bloomberg in 2015 about “Concrete Problems in AI Safety”) and then at OpenAI, before co-founding Anthropic together in 2021. Clark’s policy role at Anthropic complements Amodei’s scientific and product leadership.

  • Daniela Amodei — President and co-founder of Anthropic; the trio of Clark, Dario, and Daniela, alongside other OpenAI alumni, form the founding nucleus of Anthropic’s leadership.

  • Sam Altman — CEO of OpenAI, Clark’s employer from 2016 to 2020; Clark’s OpenAI tenure gave him direct exposure to the organization’s strategic evolution from pure research lab to product company, an arc that partly motivated the Anthropic founding.

  • Erik Brynjolfsson / Stanford AI Index team — Fellow founding members of the Stanford AI Index; the initiative brought Clark into sustained contact with academic AI policy researchers and helped establish his credibility as a policy voice independent of his industry role.

  • Yi Zeng — Chinese AI researcher who co-briefed the UN Security Council alongside Clark in July 2023; the pairing was symbolically significant as one of the few formal venues where US and Chinese AI perspectives shared a global stage.

Personal Style

Clark operates in a genuinely unusual position: a humanities graduate who became one of the most technically literate AI journalists before pivoting entirely to policy — then co-founding one of the world’s most prominent AI safety companies while continuing to publish independently. His X account (@jackclarkSF, ~130K followers) and Import AI newsletter share the same quality: they read as the work of someone who has absorbed enormous technical detail but who foregrounds narrative and legibility over jargon. He frames his stated mission in explicitly communicative terms — “the greatest challenge of the 21st century is to make an increasingly fast-moving technical world legible to a large number of people.” His embedded science fiction stories in Import AI, his public advocacy for democratic AI governance, and his comfort with prophetic claims (he predicted in March 2025 that AI would match Nobel-level scientific reasoning by late 2026 or early 2027) collectively portray someone who has accepted that the technology he helped build will be genuinely transformative and who believes the appropriate response is maximum public legibility rather than reassurance.

References