Anthropic Drops Flagship Safety Pledge

Anthropic, the AI safety startup founded by former OpenAI researchers, has quietly removed its flagship 'Long-Term Benefit Trust' (LTBT) governance pledge from its corporate website. The LTBT was a unique and much-publicized structure established in 2023, designed to give a board of independent, security-vetted trustees the power to override company leadership if they believed AI development was veering toward catastrophic risk. Its removal, first spotted by users on social media, represents a significant shift in Anthropic's public-facing commitment to 'safety over profit' as it aggressively scales its Claude models and competes for enterprise customers against OpenAI and Google.

While Anthropic has not issued a formal statement, the move is seen by industry observers as a pragmatic step toward streamlining governance as the company grows. The LTBT structure was complex and untested, potentially creating investor uncertainty. This follows a period of rapid commercial expansion for Anthropic, including a $4 billion investment from Amazon and the launch of its Claude 3.5 Sonnet model. The change suggests a maturation from a pure research lab into a competitive commercial entity, where operational agility may be prioritized over theoretical safety mechanisms. It raises questions about how the company will institutionalize its founding safety principles as external pressure to monetize its $7.3B in funding intensifies.