Jack Clark is the co-founder and Head of Policy at Anthropic, an AI research and safety company. He has been instrumental in setting strategic directions for AI policy and governance. Clark, who originally hails from Brighton, England, has a diverse background that includes roles as a technology journalist and Policy Director at OpenAI. His work focuses on ensuring that AI technologies are developed in a manner that is safe, ethical, and aligned with human values. ^{[1] [2]}​

Early Life & Education

Jack Clark was born and raised in Brighton, England. He pursued higher education at the University of East Anglia, where he majored in English literature and creative writing. ^[1] His educational background in the humanities may seem unconventional for a leader in the tech industry, but Clark credits it with providing him the analytical skills necessary to navigate the complex narratives and ethical implications surrounding AI technologies.

Career

Journalism

Jack Clark began his career as a technology journalist, covering enterprise infrastructure, distributed systems, cloud computing, artificial intelligence, and related technologies. He worked as a researcher and reporter at Adfero, reporting on technology, finance, and environmental topics. He later joined CBS Interactive (UK) as a reporter for ZDNet UK, where he covered enterprise infrastructure and interviewed executives from technology companies.

In 2013, Clark joined The Register as a reporter covering distributed systems, datacentres, high-performance computing, databases, cloud computing, computer clusters, and networking. He moved to Bloomberg LP in 2014 as an Enterprise Tech Reporter, writing about artificial intelligence, cloud computing, big data, and enterprise software. In 2015, he became a reporter covering Google, artificial intelligence, and Amazon Web Services, with reporting that included neural networks, distributed systems, evolutionary algorithms, semi-supervised learning, and machine learning.

OpenAI

Clark joined OpenAI in October 2016 in a Strategy & Communications role. In October 2018, he was appointed Policy Director. His responsibilities included communicating technical developments in artificial intelligence to policymakers, conducting research on AI evaluation, measurement, and analysis, managing an interdisciplinary team, and providing testimony before the United States Congress on matters related to artificial intelligence.

Anthropic

In January 2021, Clark co-founded Anthropic with six former OpenAI researchers. He serves as Co-Founder and Head of Policy, where his work is related to artificial intelligence policy, governance, and safety. His role includes work on oversight approaches and policy frameworks associated with AI development.

Import AI

Since August 2016, Clark has written Import AI, a weekly newsletter focused on artificial intelligence research, industry developments, and public policy. The newsletter summarizes developments in AI research and business and is distributed to readers interested in the field.

Policy and AI Governance

Clark serves as a member of the OECD working group on AI classification and definition and is a co-chair of the AI Index at Stanford University. He also participates in public discussions on artificial intelligence through interviews, publications, and Import AI, addressing topics related to AI policy, governance, and safety. ^{[12] [11] [10] [9] [8] [7] [6] [5] [4] [3] [2] [1]}

Interviews

Views on AI Policy and the Future of Artificial Intelligence #01

On September 30, 2024, Jack Clark discussed artificial intelligence (AI), public policy, and AI governance during an interview published on Anthropic's YouTube channel. Throughout the discussion, he presented his perspective on the development of AI systems, their role in society, and policy considerations related to their deployment.

According to Clark, AI systems differ from conventional software because they display patterns of reasoning and content generation that reflect information learned during training. He used analogies such as "silicon countries" to describe advanced AI systems and referred to misaligned models as comparable to "rogue states" when explaining his view that governments should address AI through coordinated policy across multiple institutions rather than through isolated regulatory measures.

Clark also discussed the difference between the speed of machine processing and human decision-making. He stated that increasingly capable AI systems may require oversight mechanisms adapted to faster operational timescales, including automated monitoring systems, usage constraints, and other governance approaches designed to manage system behavior.

The interview also addressed the relationship between AI adoption and public policy. Clark stated that AI systems could be applied to administrative processes, education, and information analysis while also presenting challenges related to misuse and system capabilities. He described these issues as subjects that policymakers would need to consider as AI technologies continue to develop.

When discussing future policy developments, Clark said that 2025 was expected to include additional international regulatory activity related to AI. He referenced the implementation of the European Union AI Act, AI safety summits, and the involvement of governments and international organizations in developing governance frameworks. Clark also stated that wider public use of AI systems could contribute to understanding their capabilities and limitations. ^[13]

AI and Public Policy #02

On June 24, 2026, Jack Clark appeared on The Reason Interview with Nick Gillespie and ReasonTV, where he discussed artificial intelligence, public policy, and the development of AI systems. During the interview, Clark described his views on the application of AI in education, administrative work, and scientific research. He said that AI systems could be used as educational tools, automate certain business and administrative processes, and assist researchers in fields including biology, medicine, mathematics, and physics.

Clark also discussed topics related to AI model behavior, including hallucinations, sycophantic responses, military applications, and recursive self-improvement. He described recursive self-improvement as a scenario in which AI systems could contribute to the development of subsequent AI models with reduced human involvement. According to Clark, this possibility raises questions about methods for evaluating, monitoring, and governing future AI systems.

The interview included a discussion of AI governance and regulation. Clark stated that decisions regarding advanced AI systems should involve governments, technology companies, researchers, and other stakeholders rather than being made by a single group. He also outlined Anthropic's publicly stated policies regarding the use of AI for mass surveillance and fully autonomous weapons, and discussed regulatory approaches including transparency requirements, third-party testing, and international standards.

Clark also commented on AI and employment. He said that available data had not established a definitive relationship between AI adoption and large-scale job displacement, while noting observations related to entry-level hiring in some sectors. He also described Anthropic's "Claude Corps" program, which places recent graduates in nonprofit organizations while providing experience with AI tools. Throughout the interview, Clark discussed AI as a technology with technical, economic, and public policy implications, and addressed topics including government oversight, international coordination, and long-term planning. ^[14]