More than fifty New Yorkers with concerns about the generative AI industry gathered in a modest auditorium on the second floor of the labyrinthian St. Michael’s Church last night for a town hall hosted by a newly launched advocacy group called Humans First. 

This Upper West Side location was chosen because it’s in the city’s 12th congressional district, where AI companies are pouring millions into efforts to stop former Palantir data scientist Alex Bores from winning an open seat. Ironically, one of Palantir’s co-founders is helping pay for attack ads that accuse Bores of profiting from Palantir’s AI surveillance tech that the company is selling to ICE. Bores says that’s why he quit Palantir and part of why he’s pushing to regulate the biggest AI companies. 

Bores wasn’t in the church last night. No candidates were. The agenda was centered around everyday people’s concerns with AI, as well as a pledge for politicians to refuse to accept money from “Big AI,” referring to a list of 11 companies including Palantir, Anthropic, OpenAI, and X, as well as people who have invested over $150 million into them. 

As you might suspect, the audience appeared to lean liberal. There were people wearing merch for socialist Mayor Zohran Mamdani. Three Humans First organizers spoke, all of whom have a background in leftist organizing. Bernie Sanders was name-dropped. Pronouns were shared. Peter Thiel was compared to the Terminator. 

One Humans First organizer who didn’t speak and who sat in the back corner of the room was Joe Allen, a former correspondent on former White House chief strategist and close Jeffrey Epstein associate Steve Bannon’s War Room podcast. Allen voted for Donald Trump, recently used the r-slur on X to describe “most” opinions about AI, and casually mentions Bannon going to jail for refusing to testify about Trump’s efforts to overturn the results of the 2020 election. He is also ardently, aggressively opposed to AI technology. 

Allen was profiled by TIME Magazine last month in its feature called “The People vs. AI.” The outlet reported that he left the War Room podcast after five years of talking to Bannon about “transhumanism,” or the idea that humans are merging with machines via AI, robots, and smartphones, to help start Humans First. The TIME article was the first thing Humans First posted on its X and Instagram accounts at the end of February.

The organization has been busy since then, hosting nine town halls this month so far. They started in the south, hitting four cities in Florida in just the first week. One place Allen promoted the tour was on War Room, where Bannon said Allen had been working behind the scenes for months with “major players” in the industry. Allen teased that “very notable figures” would be joining him on the steps of the state capitol in Tallahassee. 

Two of those figures turned out to be Alice Rothbauer and Amy Kremer. Rothbauer, a former pro-life crisis pregnancy volunteer and the current Vice Chairman of the Republican Party of Sarasota County, is the Humans First Florida State Director. Kremer is the cofounder of Women for Trump and identifies as one of the founders of the modern-day Tea Party movement. Also, the Associated Press reported that she helped organize the Jan. 6 rally, although she did not storm the Capitol. She is the Director of Conservative Coalitions for Humans First. 

The advocacy group’s focus on Florida is tied to its support of legislation proposed by Gov. Ron DeSantis called the Florida AI Bill of Rights, which cleared the state senate but will not be given a vote in the house, according to Speaker Daniel Perez. A group from Humans First, led by Kremer, showed up to Perez’s office to demand he pass the bill. Opponents of the bill, which would prohibit children from having accounts for chatbot companions without a parent’s permission, include people who oppose age verification software and digital IDs.

Apart from a brief mention of their support for the Florida AI Bill of Rights during a Q&A held at the end of the town hall, Humans First organizers stayed away from these topics with the audience in Manhattan on Tuesday, although they did stress that Humans First is nonpartisan and seeks support from across the political spectrum. Online, they aren’t trying to hide their MAGA members, whose presence is all over the official Humans First social media accounts.

When I approached Allen after the event, my number one question was how MAGA could square being anti-AI while Trump visibly cozies up to the industry. Trump remains close with Elon Musk, and the federal government under Trump has invested hundreds of millions of dollars into contracts with xAI, Anthropic (until recently), OpenAI, and Palantir. Trump is even trying to block state regulation of AI by executive order after a similar measure was defeated in the senate.

Allen told me that his support for Trump is not because of how the president handles AI. In fact, he said, Trump has been wrong on more than one occasion recently. I asked him if there's a schism forming in the MAGA base around issues like AI, and Allen said there is. (For DeSantis, who is likely considering a 2028 run, being anti-AI is one way to appeal to Republicans who have increasingly mixed feelings on Trump and J.D. Vance.) Allen also thinks there’s a similar schism on the left, with anti-establishment voters on both sides of the aisle rallying against AI at the same time. And when I asked him about the backlash to age verification laws? Well, he thinks that’s bullshit.

Therein lies some of the contradictions we see in the tech policy space. Laws that are supposed to protect minors from AI and big tech surveillance can instead surveil everyone by requiring that adults upload our government IDs for AI age verification administered by random companies to access the internet. That data has already been successfully hacked, affecting tens of thousands of people. And creating these massive databases that link internet users to their identities is an unprecedented threat to anonymity, privacy, and security online. It could be exploited by both governments and bad actors. During the New York town hall, nearly every single person in attendance raised their hand when asked if they were concerned about mass surveillance. At the same time, they’re encouraged to support legislation like Florida’s AI Bill of Rights that would likely require forms of mass surveillance to go into effect.

Policy requires a certain level of bipartisanship. But when the sides are diametrically opposed on how the world should work, that can conflict with the stated goal of legislation to, for example, “protect kids” from big tech. Take the Kids Online Safety Act (KOSA). The bill’s Republican co-sponsor, Sen. Marsha Blackburn, heavily insinuated that the true purpose of the bill is to protect kids from becoming transgender by limiting their access to information about LGBTQ people online. A Humans First organizer also touted Blackburn’s national policy framework for AI as an example of a politician working to hold the industry accountable. But is she really?

Almost exactly two years ago, I reported on another anti-big tech advocacy group backing KOSA called Mothers Against Media Addiction (MAMA). Founder Julie Scelfo, a former New York Times journalist, told me you make “strange bedfellows” in politics. During the Humans First town hall, Left Coalition Director Alexander McCoy said almost the exact same thing: he said he’s made “weird bedfellows” in the fight against big AI, including people he doesn’t agree with on practically anything else. (Interestingly, Humans First only follows 13 accounts on Instagram, and one of them is MAMA.)

In theory, these left-minded folks and their right-wing collaborators are laying down the pitchforks to unite over common issues facing everyone. But in practice, any federal laws they’re championing would likely go into effect under Trump’s administration, and state law enforcement could be up to figures like DeSantis. These Republicans have already demonstrated who benefits and who suffers from giving them more power. It takes Democrats for their bills to succeed, but the left rarely gets a say on how the laws get implemented once they’re passed.

Humans First describes itself as a grassroots organization. When I asked where their funding comes from, an organizer told me that they were originally incubated by the Center for AI Safety (CAIS). They have since gone independent and applied to become a 501(c)(3) nonprofit organization, according to their website. At the same time I was reporting on MAMA two years ago, POLITICO reported that the CAIS was spending big bucks on lobbying Congress with concerns of “severe” AI harms while simultaneously being funded by people in big tech and big AI. In fact, according to POLITICO, critics of CAIS and similar groups suggest that they’re a “smokescreen” for the biggest AI players to cement their ideal policies by warning of hyperbolic, extinction-level threats posed by AI (CAIS and the other groups disputed this to POLITICO). We hear about those supposed threats from AI CEOs a lot, and I heard it last night from Humans First organizers who suggested that AI really could supersede our intelligence and kill us all.

In the elevator up to the event, I spoke to a man who told me he got interested in attending after reading the New York Times bestseller “If Anyone Builds It, Everyone Dies,” which is a book with loose logic and reasoning about how AI will, again, probably kill us all. The book came out last September, and a few weeks later I was in D.C. While riding the metro near the halls of Congress, I noticed the train car was blanketed in ads for the book, which did not do much to dissuade from the theory that its doomsday predictions are a form of reverse psychology to lobby the U.S. government into allowing big AI to do whatever it takes to save us from extinction. And then just last week, the book’s author posted on X about a Humans First meeting that he saw happening at Harvard after he gave a talk, which he called “hopeful stuff.”

So is everyone getting hoodwinked by big AI? I wouldn’t go that far. All the town hall attendees I spoke to at length were kind, thoughtful, and empathetic about the issues AI is already creating in their communities. I talked to a 20-year-old NYU engineering student who is concerned that AI will take away entry-level job opportunities and stunt learning for young kids. I heard from a therapist whose patients are turning to AI chatbots to solve personal dilemmas, resulting in AI-generated answers that are designed to flatter them, not challenge them. And I spoke to a teacher who has started offering AI services to entrepreneurs on the side to earn a little extra money but already feels conflicted about it. I’m glad that Humans First brought us all together. I think we need more open conversations like this.

People already seem to understand and care about many of the issues presented by the rapid proliferation of generative AI tech. Creating legislation that will actually address those issues, not make them worse, is where it gets tricky. The legacy of previous laws meant to regulate internet harms, like SESTA-FOSTA, should serve as a warning of what can happen when we funnel good intentions into bad policy. But if we want those lessons to stick, we have to make sure people actually know about them. Right now, I sense most people have no clue.

Thanks for reading this far! This type of unpaywalled reporting is thanks to the generosity of Spitfire News members. You can subscribe for free or upgrade for just $5 a month to support my independent journalism.