Back in 2022, the cryptocurrency “industry” was facing some problems. Their product was sort of like money, though useful only for criminals, and also consumed nation-state amounts of electricity. A boom in NFTs—basically a receipt, except it doesn’t actually prove ownership—went bust. FTX, a major cryptocurrency exchange, turned out to be a giant fraud and collapsed. Congress and regulators at the Biden administration were closing in.
What to do? The crypto moguls put their heads together and came up with the same solution as every corrupt swindler going back centuries: just buy the government! In 2024, the crypto industry was the largest corporate political spender by far, dropping $245 million to oust critics like Ohio Sen. Sherrod Brown, and elect friendly faces like Donald Trump. With the government bought and paid for, crypto indeed got a big package of deregulation, and might get more. Trump has even gotten into the crypto grift personally.
The next Silicon Valley trend, of course, has been artificial intelligence, and the big AI companies are already following in the footsteps of their crypto predecessors by setting up some heavily funded super PACs. That’s why the progressive nonprofit Demand Progress has set up a new monitoring effort, called AI Money Watch, following one particular AI PAC (note the space), Leading the Future (LTF). The group has raised $125 million from various Big Tech villains, including the venture capital firm Andreessen Horowitz; Greg Brockman, the co-founder and president of OpenAI; and Joe Lonsdale, the co-founder of Palantir. It’s run by Josh Vlasto and Zac Moffatt, a former crypto spokesman and Republican consultant, respectively.
The Prospect obtained an exclusive preview of this effort, which will be vitally important during the upcoming midterms and in future elections.
The most obvious reason to track this spending is that any messaging the PAC produces will almost certainly be dishonest. AI as a business is quite unpopular, with 56 percent negative sentiment and just 38 percent approval in a recent NBC News poll. The data centers AI requires are even more unpopular, with a recent Heatmap News poll finding that Americans oppose them by a 71-21 margin—a 49-point swing in just one year.
When something is this unpopular, its associated PACs tend to carefully avoid mentioning what they actually care about. Instead, they run pretextual ads bringing up unrelated pseudo-objections to their enemies. That’s how crypto took down Sen. Brown, and it’s how the Israel lobby took down Reps. Jamaal Bowman (D-NY), Cori Bush (D-MO), and Thomas Massie (R-KY). So, when some ad campaign is talking about housing, jobs, or whatever, and it’s funded by LTF, it will be vitally important to point out what is really going on.
It will be even more important to identify the politicians who are taking lots of LFT money. The group’s “goal is clearly to just avoid regulation of any kind, despite what they say,” said Colin McGlynn, an AI policy adviser at Demand Progress. “They get mad when we say that because they’re like, ‘No, we support a national framework. We just don’t ever support any of the components of it that would exist. We oppose every particular proposal.’” Let’s be real: Like practically every industry in history, Big AI is against nearly any regulation unless they are the ones writing the rules.
So given that Rep. Ritchie Torres (D-NY) has taken $982,688 from LTF at time of writing (and been endorsed by LTF), it’s fair to ask what the group thought they were paying for. “If you are going to take the money from the people that say, ‘No, don’t regulate anything,’ then you’ve lost credibility,” said McGlynn.
Conversely, it is important to note that LTF has spent $3,610,624 thus far attempting to stop New York state assemblyman Alex Bores (D-NY) in the upcoming primary for the state’s 12th Congressional District seat, vacated by longtime incumbent Jerry Nadler’s decision not to run again. If they want Bores stopped, that’s a sign he must be doing something right—and sure enough, Bores is not just an AI critic, he is arguably one of the few legit AI experts in political office, having worked for tech companies, including Palantir, for years. He has since sold his stock and proposed one of the most serious AI regulatory frameworks in the country, for New York state.
That’s why they hate him. Big corporations absolutely cannot stand it when their own expert staff turns against them. They want their legislators dumb and complacent, with hands out for bribes.
I should note, as we have reported at the Prospect, that LTF is not the only AI PAC in the game at present. The other big one, Jobs and Democracy PAC, is associated with Anthropic, and since that company appears to be at least somewhat less evil, Demand Progress has decided to focus their limited resources on LTF. Incidentally, the two PACs are at daggers drawn. “You’ve never found two super PACs that hate each other more than these two,” said McGlynn.
It’s somewhat mysterious why public opinion has shifted this far, this fast, against the AI industry. At a guess, it has something to do with AI initially being sold in typical Big Tech fashion, namely by attempting to ram it down the throats of the public without any community engagement whatsoever, like one of Facebook’s innumerable hideous user interface updates, and the initial sales pitch from Silicon Valley centi-billionaires being approximately “this machine will take your job and double your power bill.” It likely also has to do with the general unpopularity of those same billionaires, and their undeniable near-death grip on the political system.
Still, even if the current AI boom turns into a flop, it is almost certainly here to stay. In some ways, my comparison to crypto above is unfair: AI is certainly proving to be a lot more useful than funny money for drug dealers and terrorists. But that arguably only strengthens the need for a new regulatory framework. Cryptocurrency could be reasonably regulated just by enforcing the laws against bank fraud and unregistered securities already on the books (and maybe repealing the pro-crypto legislation the industry purchased); AI models trained on incomprehensible quantities of socially produced data that can, for instance, solve long-standing problems in mathematics are a genuinely novel development where, say, extant copyright law is obviously insufficient.
Actually doing that might require confronting the power of the Supreme Court and overturning its Citizens United and Buckley v. Valeo rulings that let big limitless money into election campaigns. Until that happens, the least we can do is pay attention to corporate attempts to purchase the federal government.
