Anthropic Reorganizes, Launches Institute Amid Legal Battle With Pentagon

Anthropic announced a major internal reorganization Wednesday, establishing a new research division called the Anthropic Institute while its legal fight with the U.S. Department of Defense intensifies. The move coincides with significant executive changes as the AI company navigates a federal blacklisting that threatens hundreds of millions in revenue.

The newly formed institute will consolidate three existing research teams focused on societal impacts, system vulnerabilities, and economic effects. Its mandate is to examine broad questions about AI's influence on employment, safety, and human values. Co-founder Jack Clark will transition from head of public policy to lead the institute as head of public benefit. Sarah Heck, formerly head of external affairs, will now oversee a public policy team that tripled in size last year.

Clark stated the institute's creation has been planned for months, but its launch follows closely on the heels of Anthropic's lawsuit against the government. The company is challenging its designation as a supply-chain risk, a move triggered by its refusal to develop technology for mass domestic surveillance or fully autonomous weapons. This blacklist could bar Defense Department contractors from using Anthropic's systems.

"The pace of AI progress isn't slowing itself down for external events, and neither are we," Clark told The Verge. He expressed no concern over research funding, despite court filings indicating the Pentagon dispute jeopardizes at least hundreds of millions in projected 2026 income. Clark argued that rigorous safety research builds customer trust and functions as a profit center, not a cost.

The institute launches with roughly 30 staff, including notable hires from Google DeepMind, OpenAI, and academia. It plans to double in size annually and will incubate new teams studying AI's effect on sectors like the legal system. A key research focus will be the psychological relationship between users and AI, moving beyond technical analysis to large-scale social science studies on how the technology changes human behavior.

When questioned about potential conflicts if the institute's findings cast the company in a negative light, Clark pointed to Anthropic's status as a public benefit corporation and a commitment to transparency among its leadership. The company maintains that confronting AI's hardest questions is essential, even amidst commercial and legal pressures.

Source: The Verge