Dario Amodei (다리오 아모데이), CEO of Anthropic, directly cited trust issues with OpenAI CEO Sam Altman as a reason he left OpenAI.

On June 17, Business Insider reported that Amodei told a Bloomberg interview, "We do it our way, and they do it their way," adding he is satisfied with the current dynamic.

He also said he feels calm about the competitive relationship between Anthropic and OpenAI. "Why fight someone you do not share a vision with and do not trust?" he said. He added that the market and public opinion will ultimately judge the two companies' approaches. "We will see who wins in the market and who wins in the court of public opinion," he said, stressing that results matter more than the noise around the reasons for founding the company.

The remarks come as attention has returned to the background to Amodei, his younger sister Daniela Amodei, and 9 OpenAI employees leaving in 2020 to found Anthropic. A recent New Yorker report examined whether Altman is a figure who can be trusted, and cited notes in which Amodei wrote down interactions with Altman during his time at OpenAI.

Altman has also recently mentioned Anthropic directly and expressed discomfort. In an April podcast, he said apocalyptic rhetoric about artificial intelligence at some labs and the way Anthropic referred to OpenAI were not helpful. The exchange between the two companies has spilled into public remarks.

Amodei argued that even on safety issues that require cooperation among AI companies, the key is not unity itself but who sets the standards. "Trusted actors need to come together to set standards and create a situation where even people who took the opposite position have no choice but to follow the same standards," he said. He added that if much of the industry adheres to the right standards, other companies have fewer options.

He also stressed that not all AI companies distrust each other. Pointing to his relationship with Demis Hassabis, CEO of Google DeepMind, Amodei said, "I have known him for 15 years and we have dealt with various issues together." He said Anthropic buys computing resources from Google and frequently exchanges safety-related ideas. The remarks reflect a view that even within the industry, there is a divide between companies that can be trusted and those that cannot.

Amodei also referred to a moment at an India AI summit when he did not actively play along with staging industry unity alongside Altman. He said the event was extremely chaotic at the time, and that other meetings often attended by world leaders are also sometimes poorly organised. The moment has also been consumed as a symbol of the competitive dynamic between the two companies.

Against this backdrop, the interview shows that competition in generative AI is expanding beyond technical capabilities to include trust among executives, safety standards and how the industry builds solidarity. The rivalry between Anthropic and OpenAI is moving beyond product performance or market share to which company will be seen as a more trustworthy actor in AI development.