El-sayed Backs Ai Kill Switch, Safety Agency

Michigan Senate candidate Abdul El-Sayed on Tuesday sharpened his stance against advanced artificial intelligence, laying out an expansive agenda for reining in AI via a kill switch to shut off dangerous models and a new federal safety agency.

El-Sayed, who has boasted anti-Big Tech bona fides as a central pillar of his campaign, joins a growing chorus of progressives who are urging hard-line policies to slow the technology’s rapid development amid warnings from top executives about the risks AI may pose to humanity.

“The answer to what is stopping rogue AI agents from triggering a global nuclear crisis or deploying bioweapons cannot be the whims of a few billionaires in California,” El-Sayed said in a statement.

El-Sayed praised a recent bill from Reps. Ted Lieu (D-Calif.) and Nathaniel Moran (R-Texas) bill, the AI Kill Switch Act, H.R. 9917, that would give the Department of Homeland Security authority to shut down or slow AI models judged too dangerous for release.

The Michigan Democrat also re-upped his September call for a pause in AI development. On Tuesday, he also backed a ban on recursive self-improvement, a process by which AI models train successor models, and a new Cabinet-level agency that would “inspect labs, block releases, and order a pause or command a shutdown” of dangerous models.

El-Sayed’s proposal mirrors a bill by Sen. Bernie Sanders (I-Vt.) and Rep. Greg Casar (D-Texas), both of whom have endorsed El-Sayed, which would ban superintelligence and pause AI development until a government-led vetting agency could be stood up.

El-Sayed also called for global cooperation on an AI development slowdown, whistleblower protections and mandatory reporting when models escape their testing environment — as they did in a high-profile breach of AI platform Hugging Face in July.

The latest platform builds on one released in June that called for all frontier AI companies to be chartered as public benefit corporations, public ownership in the largest labs and restrictions on AI being used for sensitive decisions, like denying medical coverage or firing weapons without human approval.

With the midterms less than a month away, progressives are seeking to translate the public’s general skepticism of Big Tech into electoral victory.

POLITICO polling has shown that a plurality of Americans favor a pause in AI development and a majority think that AI could destroy humanity.

Data centers, some of which help power AI, have been a central issue in the Michigan Senate race, with Republican nominee Mike Rogers backing a moratorium on data center construction — a bold position in a party whose leaders have championed the infrastructure build-out they see as vital to maintaining American dominance in the tech industry.