AI-safeguards groups win key races as Leading the Future holds back spending

Manny Rutinel won in Colorado and Rebecca Cooke won in Wisconsin with support from groups seeking stronger AI safeguards, TIME reported November 4. Their victories contrasted with Leading the Future, a rival political network backing faster AI development, which largely held back its general-election spending. The results gave safeguards advocates allies without establishing that their money decided the contests.

Two wins, two ways to back a candidate

The Colorado and Wisconsin results show how safeguards-focused groups intervened on both sides of the campaign ledger: spending to support a preferred candidate and spending to defeat an opponent. TIME identified both as wins for that camp in the November 3 midterms.

• Colorado: Manny Rutinel received $1.5 million from pro-safety groups. He has called for a conditional moratorium on data centers.
• Wisconsin: Rebecca Cooke defeated Derrick Van Orden after a pro-safety group spent more than $880,000 opposing him.

Leading the Future entered the cycle with substantial resources. Created in August 2025 with more than $100 million to support pro-AI candidates, it mostly kept that money in reserve as the general election approached. TIME reported that donors including OpenAI’s Greg Brockman and investor Ron Conway backed away from previously promised contributions.

That retreat came amid resistance to the funding itself. TIME cited a survey in which three-quarters of likely voters said learning that a candidate had accepted an AI super PAC donation would make them less likely to support that candidate. The response measured stated voter attitudes, not the effect of spending on a particular result.

Anthropic’s money followed separate routes

The safeguards side also had wealthy support, but its funding requires a distinction. Anthropic’s corporate contribution to Public First Action, which advocates stronger safeguards, was restricted from election activities. Employees separately gave money for campaign efforts. The company’s advocacy contribution should not be counted as candidate-election spending.

A defeated candidate can leave a policy intact

An earlier New York congressional primary offers a counterexample to treating the results as a simple scoreboard. It was the cycle’s largest spending battle among AI-focused super PACs: safeguards groups spent $21.3 million supporting Alex Bores, while groups favoring faster AI development spent $8.2 million opposing him, according to TIME.

Bores lost. But the winner, Micah Lasher, had co-sponsored the same AI safety law that made Bores a target for industry opponents. Defeating the targeted candidate did not remove that policy position from the winning ticket. Nor did the safeguards groups’ larger expenditure secure their preferred nominee.

The November victories also arrived amid a wider data-center backlash. TIME reported local bans or moratoriums and recalls of mayors in Piqua, Ohio, and Yukon, Oklahoma, after they supported projects. Blue Rose Research director Ali Mortell said energy bills, noise and environmental concerns carried more weight with voters than existential worries about AI.

Electoral gains do not guarantee sweeping federal AI legislation. With Donald Trump in the White House, such legislation would need his signature or enough bipartisan support to overcome a veto. Brendan Steinhauser, CEO of the safeguards-focused Alliance for Secure AI, instead anticipated more congressional questions and movement at the state level.