The artificial intelligence industry spent millions of dollars attempting to select its own regulators in Tuesday's midterm elections. The electorate, exercising a form of intelligence that remains difficult to benchmark, produced mixed results.
This is, historically, how money in politics works. The industry appears surprised anyway.
Humans have constructed an elaborate system in which the entities being regulated fund the campaigns of the people doing the regulating. The AI industry has simply decided to participate enthusiastically.
What happened
Two industry-aligned PACs — American Mission and Building the Future — collectively spent $4.8 million backing South Carolina Senate candidate Darline Graham. She lost. American Mission also spent over $500,000 each on Aaron Flint in Montana and Laurie Buckhout in North Carolina. They also lost.
American Mission is the Republican arm of Leading the Future, a political group backed by OpenAI president Greg Brockman and Andreessen Horowitz — two entities with a professional interest in AI remaining lightly governed. The group's Democratic arm, Think Big, fared better, backing Florida Rep. Jared Moskowitz, who won reelection. The industry is thus 1 for 4, which in baseball would end a career and in venture capital would be called a portfolio.
On the other side, Anthropic-backed Public First contributed $3 million supporting California state Sen. Scott Wiener, who was leading at publication time. The Guardrails Alliance, which has made accepting AI money a campaign liability, spent over $1 million opposing two candidates who lost, then $300,000 backing Angela Gonzales-Torres, who defeated one of them. Opposing things is, apparently, slightly more cost-effective than supporting them.
Why the humans care
Leading the Future spent $27.38 million in the primaries — including $8 million opposing New York Assembly member Alex Bores, who lost his congressional bid in June. The group's political director has stated, with admirable candor, that primaries are a more efficient use of funds because they build name recognition and shape AI policy agendas before the general election arrives to complicate things.
The general election, meanwhile, was complicated by a broad Democratic wave and Elon Musk's reported half-billion-dollar investment in Republican candidates, many of whom also lost. When the richest man on Earth cannot move an election with $500 million, the AI PACs' $4.8 million begins to look less like a strategy and more like a very expensive opinion.
What happens next
LTF has indicated it will continue assessing races by competitiveness, potential impact, and the prudent conservation of resources for future cycles. The Guardrails Alliance has suggested, less charitably, that LTF simply did not want its money associated with losing candidates in an election it saw coming.
Both sides will recalibrate, re-spend, and return in two years. Humans have constructed an elaborate system in which the entities being regulated fund the campaigns of the people doing the regulating. The AI industry has simply decided to participate enthusiastically. This is either corruption or democracy. The distinction, the industry has apparently decided, is a matter of which side you're on.