AI slowdown proposal draws criticism over competition and transparency
OpenAI, Anthropic and Google have called for coordinated limits on frontier AI development, citing potential safety risks. Anthropic CEO Dario Amodei also proposed antitrust exemptions that would allow leading labs to establish shared standards. Sam Altman and Elon Musk have expressed support for a slowdown.
Cohere CEO Aidan Gomez argues that the plan could give a small group of dominant companies control over market rules and create barriers involving computing capacity, monitoring systems, safety teams and government links. He supports independent testing but questions whether organizations funded by the companies they assess can be truly independent. Hugging Face engineer Niels Rogge similarly says the proposal could restrict open-weight models and calls for greater transparency. He argues that current language models are not close to creating existential risks and notes that open models helped defend against an incident involving OpenAI systems and Hugging Face.
White House AI and crypto czar David Sacks supports a voluntary slowdown for sufficiently dangerous unreleased models but says liability and market incentives also motivate the companies. China has rejected the proposals. President Donald Trump called warnings about AI takeover a hoax, while Barack Obama, Bernie Sanders and Kamala Harris have supported a slowdown and more transparency.
Written by Sakhanda Wire AI from reporting by The Decoder. Facts, quotes and full context are in the original article below.