Anthropic CEO urges legal ban on dangerous AI
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Dario Amodei wrote an essay urging the U.S. government to legally block or deter dangerous AI deployments, arguing that current policy is too slow.
- Trump's AI executive order is deemed insufficient by Amodei, who calls for mandatory testing of AI risks including cybersecurity, biological weapons, loss of control, and automated R&D.
- Amodei proposes that frontier AI models undergo rigorous testing and auditing before release, comparing the process to safety checks for cars, airplanes, and drugs.
- Anthropic outlines an economic policy response to AI disruption, advocating for capital accounts, wage insurance, tax incentives, and an expanded social safety net, possibly universal basic income.
- Amodei recommends reforms at agencies such as the FDA to speed approval of AI‑discovered drugs and calls for a ban on domestic use of fully autonomous weapons.
- Amodei notes that public opposition to data‑center buildouts reflects broader economic anxieties about AI.
Why it matters: Amodei's call for legally binding AI testing would compel companies to undergo costly safety audits, potentially delaying deployments, while his economic proposals—wage insurance, tax incentives, and expanded safety nets—aim to protect workers from AI‑driven job loss. The policy shift could reshape tech compliance and labor protections.