US Senator Bernie Moreno accused Anthropic CEO Dario Amodei in an October 5 letter of focusing public debate about powerful AI too heavily on possible catastrophe. The Ohio Republican urged more attention to applications intended to defend against security threats.
Moreno’s criticism targets the tone of the warnings. He cites earlier statements by former Anthropic researcher Jacob Coxon, safety researcher Evan Hubinger and Amodei himself. Their numerical catastrophe estimates remain the speakers’ personal judgments, not measured probabilities.
The senator calls for AI to help secure critical infrastructure and counter cyberattacks. He connects that aim to national security and competition with China. Risks from powerful systems need addressing, he writes, but doing so should not require halting development.
The signed three-page letter records a political counterposition to recent warnings, rather than a technical safety evaluation. Its praise for signing the White House agreement concerns an earlier measure. A new public response from Amodei to this letter could not be established during the review.