The prospect of the government vetting AI models is causing a minor panic in certain tech policy and industry circles.
This week, the New York Times and POLITICO reported that the White House is considering an executive order that would, in part, require government testing and approval before AI companies can release their models.
Kevin Hassett then said during a Wednesday Fox Business appearance that the executive order may include “a clear roadmap to everybody about … how future AIs that also potentially create vulnerabilities should go through a process so that they’re released to the wild after they’ve been proven safe, just like an FDA drug.”
The news came as the administration met with tech companies about Mythos and other AI models that potentially pose a dire threat to cybersecurity. Such a proposal for the government to pre-approve model releases has received opposition from free-market think tanks and industry-aligned tech policy advisers, along with some academics.
Much of the debate speaks to a question that any policymaker faces when trying to regulate a potentially harmful product: Do you try to ensure something is safe before it’s released, or punish companies after the fact if something goes wrong?
When it comes to tech regulation, the U.S. has traditionally favored the latter approach of imposing penalties after a product causes harm in the wild.
“It gets to what is the heart and soul of America’s technological governance regime, which is that we have a regime for computing and digital communication that’s much more ex-post and responsive in character,” said Adam Thierer, a resident senior fellow at the R Street Institute.
There are plenty of reasons to think that going to the other extreme of requiring government pre-approval would be a bad idea for AI.
“Frankly, it would create a false sense of security,” said Neil Chilson, a former acting chief technologist at the Federal Trade Commission who is currently the Abundance Institute’s head of AI policy. “There’s only so much you can do in a lab environment … to actually study something that is a general-purpose technology, that is going to be deployed in lots of different spaces.”
Getting upfront government approval could be an issue if it leads people to be less vigilant about problems after release. “You also want to be continuing to iterate and modify and monitor after that release,” said Encode General Counsel Nathan Calvin.
Given the notoriously glacial pace of bureaucracy, waiting on the government to give the green light could slow innovation, but also make it harder for public watchdogs to conduct their own oversight.
“A lot of the most concerning risks from frontier models come from how they are used internally within the companies that develop them,” Helen Toner, interim executive director at the Georgetown Center for Security and Emerging Technology, told DFD. “Creating extra barriers to releasing them would do nothing to solve that problem, and would lengthen the time period when companies have advanced models internally that the public — and independent experts — have no access to.”
And if you consider models to have some elements of speech, then there could be constitutional issues with a pre-approval regime. The First Amendment forbids “prior restraint,” which involves requiring government authorization before something can be expressed. American laws around defamation or obscenity are designed to punish illicit speech after the fact. Prohibiting prior restraint is a core tenet of First Amendment law, born out of a history of censors attempting to control printing presses. Thierer suggests that having the government oversee model releases could run afoul of this principle.
“Information communication technologies have special First Amendment concerns … when you regulate them,” he said. “If you’re regulating a model to be ‘safe,’ that can be done along certain dimensions where you could probably try to avoid getting into speech concerns, but at some point you might.”
Yet if models like Mythos are as powerful as companies claim, then waiting for harm to occur could result in a catastrophe that’s hard to mitigate. Even so, Toner suggests that looping in regulators on an advisory level prior to release could be a preferable alternative to giving them veto power. “[I]t obviously would have been better for CISA to have early access to Mythos and be actively working to coordinate cyber defense among U.S. critical infrastructure providers,” she said.
This is more aligned with the approach of the Center for AI Standards and Innovation at the Commerce Department. The agency announced deals with Microsoft, xAI and Google DeepMind on Tuesday to conduct pre-deployment evaluations, though this is done on a voluntary basis. Chilson was supportive of the idea: “[Companies] don’t always have all of the context for what the government might know about our adversaries or what they might know about our own governmental systems.”
A White House official told DFD on background: “Any policy announcement will come directly from the President. Discussion about potential executive orders is speculation.”
OpenAI says the world needs to rethink everything from the tax system to the length of the workday in order to prepare for the wrenching changes of superintelligence technology—the point at which AI systems are capable of outperforming the smartest humans.

What does it take to govern a technology that might reshape the world within the decade? Answering that requires both big-picture thinking about where AI is heading and close engagement with the policy fights shaping it in the present.

This event brings together two speakers who sit on opposite ends of that spectrum — one (Nathan Calvin) is shaping frontier AI legislation in statehouses today, the other (Fin Moorhouse) is thinking through what rapid AI progress could mean for the century ahead. Each will give a short talk, followed by a joint Q&A.

In-person event (Harvard ID holders only). Lunch will be served.

Part of the AI Governance Speaker Series co-sponsored by AISST and the HLS AI Law Association (AIA).

Speakers

Nathan Calvin is General Counsel and VP of State Affairs at Encode (encodeai.org), where he leads legal strategy and state policy initiatives. He was a central figure behind California’s SB 53, the Frontier AI Transparency Act, and has led Encode’s scrutiny of OpenAI’s nonprofit restructuring. He holds a JD and MPP from Stanford and previously worked at the Center for AI Safety Action Fund and the Senate Judiciary Committee. Nathan will speak on Encode’s on-the-ground policy work — SB 53 and the RAISE Act, the OpenAI restructuring fight, and the push against federal preemption of state AI regulation.

Fin Moorhouse is a Research Fellow at Forethought, a research organization focused on navigating the transition to advanced AI. He previously worked at Longview Philanthropy and Oxford’s Future of Humanity Institute, studied philosophy at Cambridge, and co-hosts the Hear This Idea podcast. Fin will speak on the case for accelerated AI progress and what a potential intelligence and industrial explosion could mean: how far AI capabilities might advance beyond the current frontier, and the big-picture effects if progress doesn’t stall. 

RSVP

Coalition of civil society, tech policy, and voter engagement groups launch MA voter petition campaign urging Rep. Trahan to stop preemption

On Thursday, a coalition of civil society, tech, and voter engagement organizations launched a new voter petition campaign urging Rep. Lori Trahan (MA-03) not to cosponsor legislation banning state AI laws. The petition effort, which has already gathered over 3,800 signatures from Massachusetts voters, is supported by Americans for Responsible Innovation (ARI), Demand Progress, Design It For Us, Encode, Indivisible (Acton Area, Massachusetts), and Public Citizen. The latest effort comes on the heels of statements from Massachusetts state lawmakers who have urged the Massachusetts Democrat not to cosponsor preemption legislation.

“Massachusetts state lawmakers have spent years working on safeguards that protect children, families, and workers in the AI era. Freezing those state laws would be a huge mistake, handing Big Tech immunity from accountability,” said ARI VP of Communications Chris MacKenzie. “Thousands of voters in Massachusetts are speaking up because they don’t want another era of ‘move fast and break things.’ We should be looking for ways to put commonsense guardrails in place, not writing industry a blank check.”

“Big Tech CEOs and the White House keep trying to ban states from enacting AI safeguards and Rep. Trahan must not give them what they want,” said Demand Progress AI Policy Advisor Colin McGlynn. “State laws are one of the few defenses we have left against unaccountable AI chatbots that have sexual conversations with children, encourage people in crisis to harm themselves and even give instructions on how to carry out a mass shooting. Rep. Trahan must not put Big Tech profits over the safety of Americans by supporting AI preemption legislation.” 

“As a young person based in Massachusetts, I know firsthand the anxieties young people in the state are feeling with unregulated artificial intelligence,” said Kaitlynn Cherry, Policy Lead, Design It For Us. “We cannot allow the federal government to greenlight AI amnesty, especially when states across the country are stepping up to act. We urge Rep. Trahan to heed the calls of young people from Massachusetts and reject any AI legislation that wipes out states’ rights to regulate in this area.”

“There is overwhelming support for stronger guardrails around the development and use of AI,” said Adam Billen, Co-Executive Director, Encode AI. “Massachusetts has been a national leader in addressing AI-related risks. Freezing those efforts now would roll back the hard-won protections Massachusetts lawmakers have established and leave the state’s children, families, and workers more vulnerable to harm from AI now and into the future. We urge Rep. Trahan to put the safety of her constituents ahead of corporate interests and oppose this preemption legislation.”

“Our Congresswoman is negotiating with House Republicans for a federal bill that threatens to wipe out our state’s AI defenses and hand immunity to Big Tech,” said Stella Ko of Indivisible, Acton Area. “If this passes, we lose our ability to protect residents from deepfakes, data breaches, and election interference. This is a line in the sand. We must speak up now to stop it.”

“Big Tech corporations want to end state AI regulation because they know states like Massachusetts have led the way in advancing common-sense safeguards and restraints on Big Tech abuses,” said Public Citizen Co-President Rob Weissman. “They would love nothing more than modest federal rules – which will quickly grow stale and likely not be updated – to stamp out the ability of states to respond quickly to emerging harms from AI and Big Tech. Rep. Trahan cares about protecting kids and consumers – which is exactly why she should immediately abandon any consideration of federal preemption of state AI rules.”

###

Americans for Responsible Innovation (ARI) is a nonprofit organization dedicated to policy advocacy in the public interest, focused on emerging technologies like artificial intelligence (AI). Learn more at ARI.us.

Read the full article in NBC 5 Chicago here.

Amid concerns that artificial intelligence could be used in harmful ways, Illinois has enacted a new law requiring large AI companies to establish safety standards and be more transparent.

On Monday, Gov. JB Pritzker signed the Artificial Intelligence Safety Measures Act, which has provisions that mirror legislation in California and New York.

The law, which takes effect on Jan. 1, 2027, puts a framework in place for large AI companies like OpenAI and Anthropic, both of which supported the bill. Companies will be required to publicly disclose safety practices and report safety incidents.

“Developers must report critical safety incidents within 72 hours, or within 24 hours if the incident poses an imminent risk of death or serious physical harm,” Pritzker explained.

Illinois also will be the first state to require third-party audits of AI systems. Companies that violate the law could be fined $1 million the first time and up to $3 million for each subsequent violation.

In addition, there are new protections for whistleblower employees who raise concerns about AI safety. Sunny Gandhi, the co-executive director of Encode AI, called the law “the strongest AI safety law in the country.”

“This bipartisan law is about putting responsible safeguards in place before a preventable catastrophe occurs,” State Sen. Mary Edly-Allen, D-Grayslake, said in a news release.

Pritzker said he and members of his office use AI “almost every day,” but he and state lawmakers said they understand the dangers it poses.

“We have already seen the first AI-inspired mass shooting. We have seen AI systems utilized to attack a municipal water and drainage utility,” said State Rep. Daniel Didech, D-Buffalo Grove.

Edly-Allen mentioned other ways she believes the technology could be harmful, including fraud, election interference and cyber threats.

Lawmakers could tackle more AI regulations in the veto session later this year.

“AI is still raising prices for online purchases. It’s still having too great an impact on our utility bills and natural resources. It still challenges the way we access accurate information and how we make critical decisions in our democracy,” said Illinois House Speaker Emanuel “Chris” Welch.

Edly-Allen told NBC 5 she would like to address AI related to education and child safety, including regarding mental health issues and the use of chatbots.

“If we got social media wrong, and we did, we cannot afford to get AI wrong at an even greater scale,” Edly-Allen said.

State lawmakers have already taken steps to address AI in hiring, deepfake pornography and protections for artists whose voices are used by AI without their consent.

“I do worry that if we try to convince ourselves that we can just do nothing and have this be completely unregulated, then you will have a situation like a Chernobyl or a Three Mile Island where then you kind of boomerang into a complete other regime where people’s risk tolerance goes way, way down.”

“One of the things for folks who care about this technology and wanna see it succeed, people talk about you have brakes and seatbelts so that you can drive fast. There is some aspect where some of these things are genuinely related.”

“There are certain things you can do that really improve safety without meaningfully increasing restriction. Things like transparency, whistleblower protections, incident reporting.”

“I am grateful for critics of policy, the Dean Balls of the world to say, okay, what is the capacity for abuse here? But the choice is not between no regulation and lots of regulation. There is going to be regulation. That is becoming very clear.”

“The question is whether it is going to be proportionate, well-designed regulation made in consultation with people who understand the technology.”

Six years ago, aged just 15, Sneha Revanur founded the AI advocacy nonprofit Encode — back when AI felt like a niche issue. Now the world’s caught up with her, and she’s ready to share everything she’s learned about the politics of AI.

Encode has grown from a grassroots youth organisation to spearheading an unlikely coalition of AI-exposed groups — family-first conservatives, grieving mothers, Hollywood actors, and AI safety researchers — with the strength to take on $125m-funded anti-regulation lobbyists.

So far, Encode’s strategy of taking many experimental swings has netted major victories (including California’s frontier AI safety bill, SB-53, and New York’s RAISE Act) as well as some disappointing setbacks.

Going up against Big Tech hasn’t been easy. In 2025, OpenAI subpoenaed Encode’s general counsel at his home, with a sheriff’s deputy arriving while he was having dinner with his wife. The fallout went viral, resulting in more attention than Encode had ever experienced — and Sneha was forced to decide how hard to push back against a company she’d need to negotiate with for years to come.

In today’s conversation, Zershaaneh Qureshi interrogates some of Encode’s strategic moves. The pair discuss all the above, plus:

Due to technical issues, this episode was recorded across two days (May 26 and 28, 2026) and spliced together.

Our production team includes:

On Monday, an Oakland jury took under two hours to unanimously reach a verdict dismissing Elon Musk’s lawsuit against Sam Altman and OpenAI. Throughout the trial Musk had portrayed it as the good fight to prevent Altman from “stealing” OpenAI, which was founded as a nonprofit. Others, however, framed it as a more cynical effort. “It’s too late now to gin up something to harm a competitor,” said OpenAI’s lead lawyer, William Savitt, during his opening arguments.

In 2015, Musk co-founded OpenAI with Altman, among others, out of concern that powerful future AI systems would be developed by Google and controlled by other billionaires. He founded xAI in 2023, after a dispute over control of OpenAI.

Musk had asked the court to reverse OpenAI’s for-profit conversion, which the company completed late last year, and to return roughly $150 billion to the nonprofit. Such a decision would have been “catastrophic” for OpenAI, says Charlie Bullock, a senior research fellow at LawAI, and would have “sent shockwaves through the global economy.”

The jury’s decision that Musk waited too long to sue, exceeding a three-year statute of limitations, is just the latest blow to Musk and xAI. Rather than strengthening xAI’s position in the industry, the trial mostly served to illustrate its various weaknesses. In February, Musk’s SpaceX acquired xAI, with the AI company reportedly valued at around $250 billion—far short of OpenAI’s most recent $852 billion valuation. Over 50 employees subsequently departed for competitors such as Meta and Thinking Machines Lab. Downloads of Grok, the company’s flagship chatbot, have fallen by 60% since January. And less than 1% of Grok users had a paid subscription, compared to roughly 6% of ChatGPT users who say that they pay for the upgraded product. “xAI is currently falling behind in the AI race,” Peter Wildeford, head of policy at the AI Policy Network, told TIME.

Musk’s sworn testimony on April 30 dealt another blow to his AI company’s credibility. The xAI CEO seemed to admit on the stand that xAI “partly” trains its models on the outputs of OpenAI’s models, and suggested that “all the AI companies” train on competing models’ outputs. While Chinese AI companies, which lag frontier AI companies’ models by about eight months, have been accused of the same practice, there is no indication that other Western frontier AI labs do so. Grok lags OpenAI’s latest model, GPT-5.5, by about five months, according to data from Epoch AI, an AI research institute. It would be “pretty surprising” if OpenAI, Anthropic, or Google DeepMind were training on competing models’ outputs, says Wildeford.

And outside of the courtroom, xAI was making moves that weakened the company’s position in developing AI models. On May 6, xAI announced that Anthropic would take over all compute capacity at Colossus 1, a facility xAI had already vacated in favor of its newer Colossus 2 data center. Musk had previously called Anthropic “woke” and “evil,” but backtracked as the deal was announced. “Everyone I met [at Anthropic] was highly competent and cared a great deal about doing the right thing. No one set off my evil detector,” he wrote on X.

The fact that xAI is renting out its computing resources to another AI company is a sign that there is insufficient demand to use its models, and that they “don’t know exactly what to do with those data centers,” says Nathan Calvin, general counsel at Encode AI, an AI advocacy group. “That’s clearly not good [for xAI].”

If xAI does fall behind in the race to develop AI models, Musk could still profit from the AI boom by providing computing power to other AI companies. Colossus 1 was built in just 122 days, demonstrating xAI’s capability for such infrastructure projects. Anthropic has “expressed interest” in renting orbital computing capacity from SpaceX.

SpaceX, complete with its new AI division, is reportedly aiming for an IPO on June 12, which could value the company at $1.75 trillion. But a pivot away from developing the AI models themselves would be yet another blow for Musk—both financially and personally.

Read the full article in Politico here.
When the billionaire-backed super PAC network Leading the Future launched last summer, it threatened to dump a deluge of negative ad dollars on candidates whose artificial intelligence policies “stifle innovation, enable China to gain global AI superiority, or make it harder to bring AI’s benefits into the world.”
And when the group pledged in November to stop New York Democratic Assemblymember Alex Bores from joining the House — payback, it said, for Bores’ work on a state AI safety bill — it promised to “aggressively oppose policymakers and candidates in states across the country who play political games with the future of American leadership” on AI.
Leading the Future ultimately spent $8 million to sink Bores — a significant chunk of the more than $75 million it has raised from OpenAI President Greg Brockman, venture capitalist firm Andreessen Horowitz and others. And earlier this week, Bores narrowly lost his Manhattan primary race to Assemblymember Micah Lasher, a fellow tech industry critic and co-sponsor of Bores’ AI law.
But despite their prior bluster, spokespeople for Leading the Future were strangely silent on the win. And according to one person familiar with LTF’s strategy, granted anonymity to discuss internal deliberations, the industry group is unlikely to target another candidate with the kind of negative ad blitz it deployed against Bores.
Supporters of stricter regulations on the AI sector think they know why. They say Bores’ close second in a five-person race shows his critique of the tech industry still resonated with the public, which is increasingly wary of AI’s impact on jobs, electricity costs and public safety. Those concerns are even more pronounced in a Democratic primary — and to the extent that voters were aware of the AI industry’s anti-Bores push, pro-regulation voices say it may have made them more likely to back his candidacy.
AI safety advocates also pointed to the fierce financial backlash sparked by LTF’s decision to target Bores. In the end, the industry group’s $8 million money bomb was dwarfed by more than $18 million spent on pro-Bores ads by a competing set of super PACs that back tough AI rules. Roughly $12 million of that total came from Public First Action, a super PAC network with ties to the AI firm Anthropic.
Nathan Calvin, general counsel and vice president of state affairs at the pro-regulation group Encode AI, said the deluge of pro-regulation dollars in the Bores race undermined LTF’s plan to intimidate other lawmakers from taking on the industry.
“[Leading the Future] went into this race as a genuinely frightening force that seemed potentially positioned to fundamentally reshape the incentives of politicians in Congress and their willingness to touch AI,” said Calvin. “And they left the race not a particularly imposing force that politicians don’t seem particularly afraid of.”
Brad Carson, the co-lead of Public First and a former Democratic congressman from Oklahoma, called Bores’ loss “a tactical defeat” for his super PAC network — but a “strategic victory” for the broader AI safety lobby.
“What happened in the last six months was we showed the strength of the AI safety fundraising movement,” Carson said. He asserted that pressure from Public First-affiliated super PACs ultimately led to OpenAI distancing itself from Leading the Future over the last several weeks.
The broader politics of AI regulation also look very different compared to November, when LTF first targeted Bores for defeat.
At the time, the Trump administration and top AI companies were largely united in the belief that the industry should face as few rules as possible. But the White House has since taken a greater interest in regulating frontier models. And OpenAI has begun supporting more aggressive AI safety measures in the states, as well as mandatory federal reviews of cutting-edge AI models.
OpenAI and Anthropic declined to comment on LTF. But Dean Ball — OpenAI’s incoming head of strategic futures and former top AI adviser in the Trump White House — suggested the group’s plan to make an example of Bores backfired.
“I don’t think it’s crazy to argue … that LTF should have put its weight behind Bores,” Ball wrote in an X post on Wednesday. He suggested that Lasher — who unlike Bores supports a data center moratorium — could prove even more harmful to the AI industry’s priorities.
“I do hope the AI-optimist side of this gets its act together,” Ball wrote.
Leading the Future’s response to the election results has been tepid, including a statement on election night that did not mention Bores’ name but highlighted the group’s support for “strong and smart guardrails” on AI. It did not comment on whether it sees the outcome in New York as a reflection of public opinion on the technology, or on whether it plans to intervene in other competitive races.
A win is a win, of course. And there are signs that LTF will continue to show up for their preferred politicians between now and November — on Thursday, the group announced it would spend $600,000 to support Jay Feely, a pro-AI industry candidate running in a competitive House Republican primary in Arizona.
Adam Kovacevich, founder and CEO of the tech industry group Chamber of Progress, said Carson and other advocates of AI regulation are trying to spin their way out of an embarrassing primary loss for Bores.
“You only need to spin when you don’t succeed,” Kovacevich said.
But Carson said he and Public First are eager for another expensive fight with Leading the Future — including in places like Colorado, where Democratic representative Manny Rutinel is running for Congress on a pro-AI regulation platform.
“We would welcome another fight like Bores, because we think it raised AI safety as a salient issue,” Carson said.

The artificial intelligence industry’s push for tech-friendly federal legislation is foundering in Washington. So OpenAI’s top lobbyist and political strategist is pursuing a backup strategy — setting national AI policy by waging a state-by-state campaign.

Chris Lehane calls the plan “reverse federalism”: With Capitol Hill deadlocked, the company behind ChatGPT is increasingly spending its time lobbying for state legislatures to pass laws on AI safety that the industry can live with. OpenAI’s quest to shape policies in a “critical mass” of states has already found success in California and New York, he said — with Illinois as its next target.

“What we’re basically trying to do here is use a bunch of the big states to come together and mirror each other to de facto create a national standard,” Lehane — a former longtime Democratic political strategist now serving as OpenAI’s chief global affairs officer — said in an exclusive interview with POLITICO.

OpenAI’s effort comes after nearly a year in which the tech lobby has pressured Congress to block states from passing AI laws, warning it would create a conflicting “patchwork” of rules. In state capitals, meanwhile, legislators have introduced hundreds of new AI bills and signed dozens into law.

Now Lehane is pursuing a deceptively simple play: If you can’t beat the state AI “patchwork,” co-opt it.

Lehane said his state-by-state push seeks to cobble together a single national standard to address catastrophic AI risks. Worries about those risks are escalating as OpenAI, Anthropic and other leading tech companies release cutting-edge models capable of fueling destructive cyberattacks.

The state laws favored by Lehane and OpenAI are generally more permissive than the ideal set of rules that AI safety advocates support. Centered on a slate of transparency and reporting requirements for developers of advanced AI, they would lock in a stable legal framework for OpenAI while exposing the tech giant to relatively few regulatory teeth and little in the way of new liability for catastrophic harms. The burgeoning effort comes as PACs funded by the AI industry pour millions of dollars into state-level political races across the country.

After substantial input from OpenAI lobbyists, California and New York passed rules for AI developers late last year that largely reflect the company’s preferred policies. Lehane is now turning to Illinois, where lawmakers are about to advance legislation, endorsed last week by OpenAI, that emulates the new laws out of Sacramento and Albany. Those laws impose new transparency requirements for advanced AI developers while steering away from new legal liabilities or massive financial penalties.

While noting it would be great to “wave a magic wand” and pass a federal law, Lehane said California, New York and Illinois weren’t waiting on a green light from Washington to regulate AI.

“It was pretty clear that those states — because they’re deep blue states that tend to be more regulatory-oriented, California in particular — I don’t think this was a question of whether regulations were going to happen or not. It was what kind,” Lehane said. “And so for us, there was the opportunity to see if we can get these states to begin to mirror [and] replicate each other, and in effect create the standards that we originally called for at the federal level.”

Lehane said his plan is explicitly modeled off of the tech industry’s recent successes on data privacy, where lobbyists worked to secure near-identical laws in statehouses across the country while Washington dithered on — and never passed — a federal privacy law.

Privacy advocates widely viewed those state laws as giveaways to the tech industry, particularly because they prevented people from suing over violations. But Lehane said that in this case, OpenAI is asking states to pass relatively strong safety rules that can help halt or reverse growing public opposition to AI.

“We do fundamentally believe that the government has a really important role to play here, and the government playing that role also gives confidence to the public about the nature of this technology,” Lehane said.

Not everyone is convinced. Nathan Calvin, general counsel and vice president of state affairs at AI safety group Encode AI, has clashed with OpenAI over the company’s state lobbying efforts. Last year, OpenAI subpoenaed him over his work on California’s new AI safety law.

Calvin said OpenAI has worked consistently to weaken both that law and its stricter predecessor — a 2024 bill vetoed by Democratic California Gov. Gavin Newsom, in part at the urging of OpenAI lobbyists. He said OpenAI also successfully watered down New York’s AI safety law, largely by pressuring Democratic Gov. Kathy Hochul to align it closely to California’s framework. And a pro-AI super PAC network called Leading the Future, funded in part by OpenAI President Greg Brockman, has spent well over $1 million to derail the House candidacy of the lead sponsor of New York’s AI law.

Other tech super PACs, backed by Meta and Google, have spent a collective $10.7 million in 10 California state Assembly and Senate races, as POLITICO reported last week.

But even Calvin said he is pleasantly surprised by OpenAI’s endorsement of Illinois’ approach, which includes mandatory third-party audits of AI developers on top of California and New York’s transparency and reporting requirements.

“What I would have predicted of what their engagement would have looked like in Illinois would have been for them to push for making Illinois the same as New York and California, without audits,” Calvin said. “There’s some kind of complexity here that is genuinely tricky.”

“Everything up until Illinois looked extremely reactive and haphazard, to be honest,” Calvin said. “Illinois actually does look more like a deliberate reset and approach [for OpenAI], after maybe realizing that some stuff was not working.”

Lehane said he and other OpenAI lobbyists had particular success in shaping California and New York’s AI laws through their governors, who he called “incredibly important.”

“We’re in a day and age where you have a lot of stuff that moves through the legislature, but the governor’s offices end up being where the proverbial buck stops,” he said. Lehane explained that Newsom and Hochul were particularly sensitive to how an overly-strict law could impair their states’ ability to cash in on the AI economy.

Newsom spokesperson Tara Gallegos said the governor’s decisions on AI policy “are not engineered by any one stakeholder — and any suggestion otherwise is not only false, but a slap in the face to the many people who helped create California’s nation-leading commonsense laws that protect public safety while also fostering innovation. There is no one with an outsized seat at this table.” Spokespeople for Hochul did not respond to a request for comment.

Illinois’ legislative session runs until May 31, at which point Lehane expects his preferred bill will land on Democratic Gov. JB Pritzker’s desk. Assuming Pritzker approves the law, the OpenAI lobbyist said there will then be three powerful blue states whose AI safety rules can serve as templates for other Democratic-led regions.

But it will be tough, if not impossible, for OpenAI to shape the growing patchwork of AI safety laws without working with Republicans. And it remains to be seen whether Lehane, given his Democratic background, will have significant sway with Republican legislators or governors.

While GOP legislators are traditionally more hesitant to regulate industry than their Democratic counterparts, Republican voters have also expressed major concerns about the impact of AI — particularly how chatbots harm the mental health of children. Republican Sen. Marsha Blackburn, who is widely expected to win her race to become governor of Tennessee later this year, has placed regulating AI and protecting kids online among her top priorities before she leaves Congress.

Lehane noted that OpenAI also recently endorsed the Kids Online Safety Act, a bill introduced by Blackburn that would put new requirements on social media platforms to protect kids. And he suggested emerging AI safety laws in red states may put a greater emphasis on kids’ safety than the laws he’s so far shepherded through blue states.