Anthropic Launches The Anthropic Institute to Confront Frontier AI Societal Risks

Anthropic, a leading AI safety and research company, announced the launch of The Anthropic Institute on March 11, 2026. This new initiative consolidates efforts across the company to address the profound challenges posed by powerful AI systems to societies worldwide. Led by co-founder Jack Clark in his new role as Head of Public Benefit, the institute aims to provide transparent insights from frontier AI development, helping researchers, policymakers, and the public navigate the transition to transformative AI technologies.

The institute focuses on critical areas including AI's reshaping of jobs and economies, opportunities for societal resilience, threats from misuse, the expressed values of AI systems and how society can influence them, and governance challenges like recursive self-improvement. It draws on unique access to information from building frontier models like Claude to report candidly on risks and behaviors observed during development. By stress-testing AI limits, studying real-world usage, tracking economic impacts, and incubating new research teams on forecasting and legal interactions, the institute positions itself at the forefront of AI safety research.

Key leadership includes high-profile hires such as Matt Botvinick, formerly of Google DeepMind and Yale Law School, to lead AI and rule-of-law efforts; Anton Korinek, economics professor on leave from the University of Virginia, for economic transformation studies; and Zoë Hitzig, ex-OpenAI researcher, to link economics to model training. An interdisciplinary team of machine learning engineers, economists, and social scientists will drive the agenda, with ongoing recruitment to expand analytical capabilities.

The institute builds on existing teams like the Frontier Red Team, which probes AI cybersecurity risks such as vulnerability scanning and exploit testing; Societal Impacts, analyzing user interactions and AI agent autonomy; and Economic Research, which recently published the Economic Index on AI-driven business automation. These efforts aim to understand AI behavior in the wild and ensure human oversight in advanced development scenarios.

In parallel, Anthropic is expanding its Public Policy organization under Sarah Heck, opening a Washington D.C. office this spring to shape priorities in model safety, transparency, energy, infrastructure, export controls, and global governance. The institute emphasizes a two-way engagement, partnering with affected workers, industries, and communities to mitigate risks and harness AI's potential upsides in science, security, economic development, and human agency.