ANTHROPIC
Anthropic wants to study the future before it arrives
Anthropic has launched The Anthropic Institute, a new group focused on studying the biggest societal questions linked to advanced AI.
The company says the institute will bring together research from across Anthropic and share insights that could help researchers, policymakers, and the public prepare for more powerful AI systems.
Anthropic says AI has advanced quickly over the past five years.
It took the company two years to release its first commercial model, and another three to build systems that can find serious cybersecurity flaws, handle a range of work tasks, and even support AI development itself.
The company expects progress to move even faster over the next two years.
Anthropic says this could bring much more capable AI sooner than many expect, raising bigger questions around jobs, the economy, law, governance, and the values built into these systems.
The Institute will share what Anthropic is learning as it builds frontier AI models and work with outside groups to better understand the risks and wider effects of the technology.
The group will be led by Anthropic co-founder Jack Clark, who is also taking on the new role of Head of Public Benefit.
It brings together teams focused on AI safety testing, real-world AI use, and the economic impact of AI. It is also starting new work on AI forecasting and legal issues.
Anthropic says the Institute will draw on insights from building frontier AI models, while also speaking with workers, industries, and communities likely to be affected by AI-driven change.
In brief:
Anthropic has launched a new institute to study the wider impact of advanced AI.
The company says the move reflects how quickly AI is progressing.
It is also expanding its public policy team and opening a Washington, DC office.
Bigger than the lab
The company also announced a few early hires.
Matt Botvinick will lead work on AI and the rule of law, Anton Korinek will study how advanced AI could reshape economic activity, and Zoë Hitzig will link economics research to model training and development.
Alongside this, Anthropic is also expanding its Public Policy team.
The company says that work will focus on areas such as model safety, transparency, infrastructure, export controls, and AI governance.
The team will be led by Sarah Heck, Anthropic’s Head of External Affairs.
Anthropic also says it will open its first office in Washington, DC this spring as it grows its policy presence.
It’s like opening Jurassic Park and then launching a dinosaur safety initiative - MV