Trump just explained why an AI agency cannot work. He is right

Stavros Gadinis is a professor of law at the University of California, Berkeley, and Faculty Director of the Berkeley Center for Law and Business. His book Corporate Ordering: How Corporations Navigate Social Conflict is forthcoming from Cambridge University Press in October.

On Monday the President posted that the only guardrail artificial intelligence needs is “a STRONG AND SMART (High IQ!) PRESIDENT.” The reflex in Washington and in the labs is to treat this as an obstacle. It is better understood as a warning. Since the Supreme Court decided Trump v. Slaughter in June, every federal agency save the Federal Reserve answers to the White House. An AI regulator would be no exception. Bernie Sanders wants a pause until a new cabinet agency is in place. John Thune wants an office with the power to block a model’s release, an implementer wearing an oversight badge. Treasury wants a FINRA for AI that reports to a Treasury official. Each proposal ends with a federal official that works for the President, and whom the President can fire at will. 

The labs know this. That is why Dario Amodei, Sam Altman and Elon Musk have each said, in their own registers, that they want rules and fear the regulator. They have a better option than they realize, and it is sitting in their own filing cabinets.

Every position in this debate assumes that governance begins when government acts. The record says the opposite. Before releasing AlphaFold, DeepMind consulted more than thirty outside experts in biology, biosecurity, bioethics and human rights, and their advice shaped the release. Every major lab now publishes a safety framework naming the capabilities that would stop a launch, and since January, California requires the largest of them to publish that framework and to report serious incidents within 15 days.

Every model ships with a system card,........

© Fortune