Anthropic CEO calls for slower AI development

Anthropic CEO Dario Amodei has called for slowing the pace of artificial intelligence development and strengthening control over its trajectory. In a letter published on his personal website, he stressed the need to regulate technological progress to limit potential risks.

Amodei’s appeal came days after researcher Jacob Coxon publicly resigned from the pretraining research department at the California-based company. Coxon accused both Anthropic and OpenAI of acting irresponsibly by racing to develop a self-improving superintelligence that could threaten human lives.

Coxon, who previously worked at OpenAI, told the press that the models being developed could become uncontrollable by the end of next year. He highlighted their potential to carry out rapid cyberattacks and seize power and resources, warning that they could cause existential catastrophes by the end of the decade.

Growing global concern has also renewed attention to warnings from the late astrophysicist Stephen Hawking. He previously cautioned that fully developed artificial intelligence could bring about the end of humanity and replace it entirely, as machines could redesign themselves at an accelerating pace beyond humans’ ability to keep up.

These developments reflect mounting concerns among researchers and experts about the use of AI in dangerous fields, including biological weapons, amid growing calls for binding regulatory frameworks for companies developing these technologies.