Would a Former Palantir Employee Make the Best AI Regulator in Congress?
An interview with Alex Bores, the candidate Big Tech is spending millions to defeat
Would you elect a former Palantir employee to Congress? If your gut reaction is no, you’re far from alone — and some of Silicon Valley’s richest, most influential power players, including Palantir’s own co-founder, couldn’t agree more.
The candidate in question is Alex Bores, a sitting New York State Assemblymember and Democrat running for Congress in a crowded open-seat primary. The field also includes Kennedy heir and long-time online public figure Jack Schlossberg, political commentator George Conway, and fellow state Assemblyman Micah Lasher.
At 35, Bores holds a master’s degree in computer science and cut his early professional teeth in Big Tech, specifically at Palantir, before pivoting to politics and winning his state assembly seat in 2022. But his industry background doesn’t translate to a free-pass approach to Big Tech. Bores is an outspoken advocate for rigorous, proactive AI regulation, and co-sponsored New York’s landmark RAISE Act — legislation signed into law in 2025 that requires major AI firms to develop and publish public safety protocols for their high-capacity models, among other critical guardrails.
Bores’ tough stance on AI has made him a top target for Big Tech leadership. In late 2025, Leading the Future — a super PAC funded by OpenAI’s Greg Brockman, Palantir co-founder Joe Lonsdale, and venture capital giant Andreessen Horowitz, among others — launched an aggressive campaign to block Bores from advancing past the primary. The group has repeatedly taken issue with Bores’ regulatory approach, telling WIRED in a prior statement that it is “ideological and politically motivated legislation that would handcuff not only New York’s, but the entire country’s, ability to lead on AI jobs and innovation.”
I sat down with Bores in early April, roughly 10 weeks ahead of what is expected to be a decisive primary in New York’s reliably blue 12th Congressional District. We discussed his time at Palantir, why so few lawmakers grasp the technology they’re tasked with overseeing, and what it feels like to be on the receiving end of super PAC-funded attack ads, texts, and mailers. This conversation has been edited for length and clarity.
Katie Drummond, WIRED: Welcome, Alex. Let’s start with your tech background. I found it so fascinating that you worked at Palantir — WIRED has covered the company for years, and a few months back one of our reporters had a great idea for a story: what does Palantir actually do, anyway? Most people don’t have a clear answer, and the wildest part of the story was that even some former Palantir employees couldn’t clearly explain the company’s core work. So as a former Palantir employee yourself, what’s your best explanation of what the company actually does?
Alex Bores: Palantir helps organizations make better use of the data they already have access to. It simplifies tracking changes to that data over time, speeds up data integration, and layers what’s called an ontology — a standardized framework for how data should be structured — on top of raw datasets.
The easiest way to explain an ontology is with a project I led at Palantir for the Department of Justice, where we were investigating big banks’ role in the 2008 Great Recession. We wanted to prove that banks knew the mortgages they were packaging into securities were subpar, didn’t meet their own stated standards.
One clear pattern that would prove that knowledge is if you saw a bad loan added to one security, pulled right before the security was issued, then slipped into another security that had the same baseline standards. That pattern would confirm the bank knew the loan was flawed.
The problem at the time was that existing e-discovery tools were built just to help lawyers read through documents. All the data was theoretically there: you had spreadsheets full of individual loan records, but lawyers were just told to read through them. No human can parse thousands of loans and track their movement across multiple datasets.
Drummond: Of course not, that’s impossible at scale.
Bores: Exactly. We realized the most important entity here was the individual loan itself — it was an object that needed to be tracked end-to-end. That’s exactly what an ontology lets you do. We built a system that could track individual loans, search for loans that moved between datasets, and we found hundreds of examples of that exact pattern: banks would catch a bad loan, pull it from one security, then sneak it into another. Because we found all those patterns, we helped recover $20 billion for taxpayers from bank settlements.
Drummond: What drew you to work at Palantir in the first place?
Bores: I got my master’s in computer science later in my career; my undergrad was actually in industrial and labor relations. I grew up joining my dad on union picket lines, I studied labor organizing in college, and I led a campaign against Nike when they laid off 1,800 workers without paying the legally required severance. We ended up winning that fight.
But halfway through the campaign, another student said to me: “Why do you care so much about these jobs anyway? They’re just going to get automated anyway.” That question stuck with me. We need to make technology work for people, not the other way around. I’m a Democrat; I believe government can and should be a force for good, but that means we have to prove it works.
I was looking for roles where I could help government deliver on its promises, serve people better, and figure out how to align tech with public good, not the other way around. That’s what drew me to Palantir.
Drummond: Palantir is notoriously controversial right now for its government work, particularly with the Department of Defense. You’ve said you resigned from Palantir after the company signed an ICE contract during the first Trump administration. Palantir had worked with ICE as far back as 2011, so walk me through that moment: when did you decide you couldn’t stay?
Bores: To be clear, I never worked on that ICE contract. Palantir had already partnered with Homeland Security Investigations, a division of ICE, during the Obama administration. That work focused on drug trafficking, human trafficking, and counterfeiting — work that’s not controversial, most people support it. When Trump took office in 2017, he pushed to change the nature of that work across the federal government, including at the Department of Justice, where they tried to force our team to work on civil immigration deportation cases.
As the project lead, I was able to say no. Our DOJ contract was structured into three mutually agreed case types, so we could explicitly exclude that work. But when it came to ICE, leadership made a different call. Trump pushed for ERO, the Enforcement and Removal Operations division of ICE, to get access to Palantir software to carry out mass deportations.
We asked leadership to add the same contractual guardrails I’d negotiated at DOJ: a clear clause that the software wouldn’t be used for deportations. Leadership made it clear they wouldn’t do that — they planned to renew the contract without any of those guardrails. That’s when I made the decision to leave.
Drummond: You could have taken another high-paying tech job after leaving. Why pivot to politics?
Bores: Government work was always the core of what I cared about, even if that didn’t mean politics at first. Fixing how government works is what I’ve focused on my entire career. After I left Palantir, I joined a startup that did anti-money laundering and counter-terrorist financing work, then another that helped state and local governments distribute Covid relief.
When my current state assembly seat opened up, a friend gave me a push: “You’re always fixing bad policy after the fact with tech. This is your chance to go upstream and get it right the first time. You don’t know if you’ll win, but if you win and hate it, if politics is just all the mudslinging people complain about and you can’t get anything done, you can quit in two or four years and go back to what you were doing. But you won’t get this open-seat opportunity again. This is your moment.”
Drummond: That’s great advice from a good friend.
Bores: Really good advice. So I ran, I won, and it’s been even better than I expected.
Drummond: Why’s that?
Bores: You actually can get things done if you put your head down, ignore the noise, build coalitions. Eventually, they run out of reasons not to pass your bill. I’ve passed 30 bills in my time in the state legislature — that’s about the same number the entire U.S. Congress passed in 2023. The nonpartisan Center for Effective Lawmaking, which ranks legislators across the country, named me the most effective first-term legislator from New York City. I’ve actually been able to get things done and improve the lives of my neighbors.
Drummond: You said back in 2022 that “at least one person in Albany should know how tech works.” You know far more about tech than the average lawmaker. But why do so few lawmakers understand technology, or the companies building these tools? I don’t buy it’s just an age thing — you can learn tech at any age.
Bores: 100% agree. You can understand tech no matter how old you are, and you can completely miss it even if you’re 35 like me.
Congress is dominated by lawyers — I love my lawyer friends, but we need a diversity of professional backgrounds in office. The skill set of a software engineer just overlaps less with traditional congressional work than a lawyer’s skill set does. We need people who work in multiple spaces, and tech is also this rapidly changing new field that many lawmakers haven’t had exposure to.
I got my master’s in computer science with a specialization in machine learning while I was working. When I was elected in 2022, I became the first Democrat elected to any state office in New York with a computer science degree. If I win this congressional race, I’ll only be the second Democrat in Congress with a computer science degree. There are two Republican members with CS degrees, out of 435 total members of Congress.
Drummond: That’s genuinely shocking.
Bores: Having less than 1% of congressional representatives with that background for a field this fast-moving and this important is definitely not the right balance.
Drummond: Let’s talk AI regulation. Several states have moved forward with their own laws, including New York, where you co-sponsored the RAISE Act, the Responsible AI Safety and Education Act. In short, it requires big AI developers to publish their safety testing practices. How does it work, exactly?
Bores: It only applies to the very largest AI developers. We set thresholds based on model complexity and company revenue: they have to hit at least $500 million a year in revenue. Right now, that means we’re talking about OpenAI, Anthropic, Google, and Meta.
Drummond: That seems like the right group to target.
Bores: Exactly. The law requires them to create and publish a public safety plan, and actually stick to it. You can amend your plan over time if you change your practices, but you can’t just ignore your own plan and amend it after the fact to cover missteps. They also have to disclose critical safety incidents to the state government, and we set a very high bar for what counts as a critical incident to avoid unnecessary red tape.
It also sets up a dedicated new AI office within New York state government that collects data on AI development, recommends new rules, and submits an annual report to the legislature on what policy updates are needed to keep New Yorkers safe.
Drummond: President Trump signed an executive order last year targeting state AI laws that don’t align with federal policy. I’d argue federal AI policy basically doesn’t exist right now in any meaningful form. How does that executive order intersect with the RAISE Act?
Bores: His executive order was directly targeting the RAISE Act, and other similar bills around the country like California’s SB 53. It was explicitly designed to pressure New York, California, and the roughly seven states working on frontier AI regulation last year to abandon their bills and not implement them.
It was just a punishment tactic: if you pass a regulation we don’t like, we’ll take away your BEAD funding, the funding for expanding broadband access.
Drummond: That’s a pretty classy threat — going after rural broadband to stop AI regulation.
Bores: It’s explicitly meant to hurt people in rural areas who don’t have internet access. They talked about cutting other funding too, and instructed the attorney general to sue states. It’s the first time I’ve heard anyone argue that more lawsuits against states will lead to more AI innovation.
It wasn’t based on any serious policy. It was just a gift to Trump’s mega-donors who want zero AI regulation whatsoever. And those are the exact same mega-donors funding Leading the Future, the super PAC attacking me now.
Drummond: Let’s talk about those attack ads. They launched right after the RAISE Act was signed into law, and they’re spending millions of dollars to defeat you. What was your first reaction when you saw those ads?
Bores: That they’re desperate. They know they’re on the wrong side of public opinion on this issue.
They run these hyperbolic, ridiculous ads because they know I’m the biggest threat to their push for unbridled control over American workers, our education system, and our climate policy. They’ve actually done more to highlight the stakes of this race than I ever could. Just this week, I was on a call with a tenant leader about housing policy — nothing to do with AI — and she told me she started paying attention to my campaign because of all their attack ads.
Drummond: It’s kind of a gift, isn’t it?
Bores: They’ve been amazing at raising awareness about AI regulation and AI safety. My original plan for this race was to only talk about tech and AI 5 to 10% of the time — if you care about that issue, you’re already voting for me. I wanted to focus on health care, housing, transportation, all the other core issues voters care about. But they’ve made this a central issue, and way more voters are paying attention now. That said, it’s not exactly fun. I’m a voter in this district too, so I’m on their contact list. I get their attack texts every single day.
Drummond: You get nasty texts about yourself sent right to your phone?
Bores: I do! I get attack mailers sent to my house, and I can’t throw them away — I need to see what they’re saying about me. I’ll pull the attack mailer out of my lobby mailbox, then get in the elevator with my neighbors holding this thing that says all these awful things about me. It’s a totally surreal experience.
Drummond: That’s next-level, I could never do that. Mean things on social media are one thing, getting physical mail attacking you is another.
Bores: But their whole strategy is to defeat me, because if I win, it undermines their threats to every other lawmaker considering AI regulation. Part of their goal is to intimidate other legislators. It’s not just about my race — I talk to members of New York’s congressional delegation and other lawmakers around the country who are watching this race, and I hope they don’t get cold feet. That’s exactly what the super PAC wants: for them to step back and think, “This is too risky, let someone else lead on this.”
Drummond: What do you want to see from federal AI regulation? You released a detailed framework a couple months back, right?
Bores: I put out an 8-topic, 43-subpoint framework a couple months ago because I thought it was important to be explicit about what I want to do. It covers everything from age verification to protect kids, to a comprehensive federal data privacy bill — we’re 20 years late on that, and AI makes it even more urgent because it can de-anonymize data that was previously thought to be anonymous — to labor protections, to catastrophic risk mitigation, to technical standards to stop harmful deepfakes. I got really specific, a little nerdy even.
Drummond: Sounds like a former tech guy. What was the reaction?
Bores: I didn’t know how it would land — a lot of politicians only stick to one or two controversial talking points, and here I had 43. But I was really blown away by the reception. Progressives on the left have said this is exactly the agenda we should be pushing for. Even OpenAI’s chief futurist quote-tweeted it and said while he had quibbles, it was a thoughtful, detailed plan.
Drummond: Did they tag Greg Brockman, who funds the super PAC attacking you?
Bores: I asked them to, no luck. For context, it gets even weirder: Chris Lehane, OpenAI’s top policy executive who helped launch the Leading the Future super PAC attacking me, wrote on his own blog that other states should copy the RAISE Act.
Drummond: No one can figure OpenAI out, honestly. It’s impossible to keep up.
Bores: A lot of the engineers who work there are pro-regulation. It’s really just the top executives and a handful of donors making these really confusing decisions.
Drummond: Let’s say you win this election, and we’ve got President Trump back in the White House, and you’re surrounded by colleagues with almost no tech literacy. How do you actually get any AI regulation done over the next couple years?
Bores: I’m actually pretty optimistic about bipartisan support on this issue. I agree with Josh Hawley on basically nothing, except that AI needs serious regulation. I did a talk with Marsha Blackburn, another person I disagree with on almost everything, and we saw eye to eye on this. Poll after poll shows voters from both parties want reasonable guardrails for AI — especially when it comes to protecting kids and protecting workers. We’ve already seen this play out at the state level: the RAISE Act passed with bipartisan majorities in both chambers of New York’s legislature, and the final vote was nearly unanimous — only one no vote. So there’s already common ground.
Drummond: The most common criticism of AI regulation is that it will kill American innovation and hurt our competition with China. How do you respond to that argument?
Bores: First off, the Chinese Communist Party cracks down on AI way more harshly than any proposal being discussed in the U.S. or the West. Regulation isn’t
