Anthropic CEO Says He Left OpenAI Over a Leader He Could Not Trust

Dario Amodei, the head of Anthropic, has for the first time spoken candidly about why he stepped away from OpenAI in late 2020. By his account, the decision came down to not wanting to keep clashing with someone whose vision he did not share and whose trust he could not rely on. The remark has been widely read as a pointed reference to OpenAI's current chief executive, Sam Altman.

Amodei did not leave alone. He took his sister Daniela and nine other colleagues with him to build Anthropic from the ground up. The young company has since made remarkable progress, eventually overtaking OpenAI in valuation.

Trust as the foundation of the AI industry

Amodei argued that trustworthy AI companies ought to work together to set common standards and hold less reliable rivals to them. He framed this kind of cooperation as a way to safeguard the responsible development of the entire field.

He pointed to his 15-year positive relationship with Google DeepMind's Demis Hassabis as proof that collaboration grounded in trust is possible.

That warmth stands in sharp contrast to the fierce rivalry now playing out between Anthropic and OpenAI. Recent friction was underscored when Anthropic ran a Super Bowl advertisement that took a jab at its competitor.

Amodei's comments suggest that the dividing lines between leading AI labs are not only about technology or business, but also about governance philosophy and how much the people running them can trust one another.