Anthropic CEO Dario Amodei faced pushback from other executives over his AI safety warnings at the White House on Tuesday. While companies agreed on new safety principles, peers questioned Amodei about his extreme views. Amodei defended his position, saying, "the industry should be honest with the public about risks."
AI chief executives agreed on safety principles at the White House, but privately questioned Anthropic CEO Dario Amodei over his warnings about the risks posed by the technology.
By CNBCTV18.com
AI executives presented a united front at the White House on Tuesday as major technology companies agreed to a set of principles aimed at keeping artificial-intelligence models safe. Behind the public show of unity, though, some of the industry’s biggest names privately questioned Anthropic CEO Dario Amodei over his warnings about the dangers of AI.
The discussion took place after lunch in the White House’s East Room attended by about two dozen CEOs. People familiar with the matter told The Wall Street Journal that executives, including Nvidia CEO Jensen Huang, questioned Amodei about why he had been so extreme about AI capabilities and potential risks.
Amodei defended his position, saying the industry should be honest with the public about what its models can do and should not downplay the risks.
AI CEOs agree on safety principles
President Donald Trump and House Speaker Mike Johnson brought the AI executives together at the White House, where the companies agreed to a set of principles focused on AI safety.
The discussions focused on preventing AI-powered cyberattacks, following concerns about the security risks posed by increasingly capable models. The principles include stronger internal reviews of models during development and external testing.
The agreement was presented publicly as an effort by the industry and government to work together on AI safety through model evaluations and internal controls.
Jay Clayton, director of national intelligence, described the meeting as one of the most constructive and collegial industry meetings he had attended. He said the central question was what would be best for the American people and the world.
Amodei has been vocal about AI risks
Amodei has previously warned that increasingly capable AI models could create serious risks. He has also criticised what he sees as attempts by the White House and other technology companies to play down those concerns.
His warnings have included the possibility of AI systems carrying out cyberattacks and causing significant job losses as the technology becomes more capable.
On Tuesday, however, he said that the industry could ‘win safely,’ if technology companies and the administration worked together. Trump also praised the Anthropic CEO, calling him “great”. The two had their first face-to-face meeting at a White House dinner on Sunday.
Tensions over AI self-regulation
The private questioning of Amodei comes amid differences within the technology industry over how AI safety should be handled. Anthropic, OpenAI and Google had earlier aligned around a proposal for an industry self-regulatory organisation similar to the Financial Industry Regulatory Authority, which oversees brokerage firms.
The proposed structure would have included stronger reviews of AI models during development and external testing. But Huang, Meta CEO Mark Zuckerberg and SpaceX CEO Elon Musk raised concerns with Trump that such an organisation could give too much power to leading AI companies.
The proposal was subsequently dropped, according to The Wall Street Journal report. Some of the same safety measures later appeared in the principles agreed upon at the White House.