AI groups spend $20 million in New York race pitting Bores, Lasher,

AI companies have poured more than $20 million into a New York Democratic primary race that could shape what artificial intelligence policies the federal government ultimately adopts. The race in a Manhattan congressional district pits state Assemblyman and AI safety advocate Alex Bores against fellow Assemblyman Micah Lasher and John F. Kennedy grandson Jack Schlossberg.

Two major super PACs affiliated with AI companies are facing off in New York’s 12 congressional district — the only congressional race so far where both groups are involved.

Leading the Future, whose backers include venture capital firm Andreessen Horowitz, Open AI co-founder Greg Brockman, and AI software company Perplexity, spent $8 million opposing Bores, who was a driving force in the state passing legislation requiring safety and security regulation for powerful AI models.

Countering that spending is Public First Action, which has gotten $20 million from Anthropic. The group has supported Bores to the tune of $11 million, according to Federal Election Commission data reviewed by CNBC on Monday.

Public First Action is an arm of Americans for Responsible Innovation, an AI safety group. Its president, Brad Carson, said while the group doesn’t reveal its donors, it has gotten support from employees at major AI companies he described as “mid-level people who are very scared about where the technology is going.”

The massive spending in one House race has become a proxy battle for the future of AI regulation in the U.S. and how heavy of a hand government should take as the industry grows and AI gains a broader foothold in society.

Leading the Future supports lighter guardrails around the burgeoning AI industry than Public First Action does.

Leading the Future’s co-leader Josh Vlasto, said in a statement to CNBC that the PAC “supports passing a national regulatory framework for AI that creates jobs for American workers, helps America win the race against China, and includes strong guardrails that protects the safety of kids, users, and communities.”

When the PAC was founded in August, it said it would “oppose policies that stifle innovation, enable China to gain global AI superiority, or make it harder to bring AI’s benefits into the world.”

Meanwhile, Public First Action is advocating for more restrictions not only on the outcomes of AI models, but also on how they are created.

“Safety should be designed into the AI models,” Carson said. “Regulating the outputs long, long, long after said problem has arisen does very little justice to the people who are harmed by the AI.”

That aligns with Bores’ view that while AI can be a positive development, it needs to have limits.

“Regulation is not going to be the reason we win or lose this race versus China,” Bores, an engineer and computer scientist who previously worked at Palantir, told CNBC Monday as he campaigned outside a subway stop. “We can invest in AI that’s meant to help doctors diagnose disease without encouraging the AI…