Staff Reports
Anthropic CEO Dario Amodei told CBS in a recent interview that he finds it “very strange” and even “uncomfortable” that powerful AI is being built by private companies. He said he might consider handing control to “the right combination of governments,” or at least some form of joint governance. That idea sounds noble until you think about what handing over the keys to the kingdom really means.
Anthropic CEO’s troubling suggestion
Amodei’s comment is a newsworthy development because it came straight from the head of one of the leading AI firms. He did not call for strict oversight or better audits only; he floated the notion of government stewardship. That is a big shift in tone from tech leaders who usually argue for private innovation with light-touch rules. Saying you’re “uncomfortable” with private companies building AI and that governments might be preferable is a provocative claim — and one that deserves pushback.
Markets are messy, but still safer than a Ministry of AI
Make no mistake: private companies are not saints. They chase profit, make mistakes, and sometimes act badly. But markets offer checks that governments do not. Customers walk away, investors punish failure, rivals pick up the pieces. A bad product can mean bankruptcy. A bad bureaucracy can mean permanent power. If you want incentives to fix problems fast, you want competition, not a new federal monopoly with its own vendors and sweetheart contracts.
History and common sense argue against handing AI to government
Government power tends to be sticky. Once a state controls a tool that can watch, influence, or punish, it rarely gives it up. History has plenty of examples where governments used technologies to spy on, oppress, or simply bungle things at massive scale. The idea that a “combination” of governments would be better sounds like a polite version of global control — and polite global control is still concentrated power. Concentrated power, whether in one company or one state, is a problem. But one is easier to replace than the other.
Conservative plan: fix the problem without surrendering freedom
Conservatives should push for smart, market-friendly solutions: clear liability rules so companies face real consequences for dangerous systems; narrow, targeted rules for especially risky AI uses; stronger whistleblower protections; mandatory safety tests and certification for high-risk models; and robust congressional oversight rather than handing control to a new unelected authority. International cooperation among democracies on export controls and standards makes sense — but stop short of creating a single global regulator. Private companies can and should be reined in. Just don’t trade liberty for the illusion of safety.
