President Donald Trump is retreating from plans for a $1.8 billion “Anti-Weaponization Fund” after a fierce backlash from fellow Republicans, according to two people granted anonymity to describe the decision ahead of a public announcement.

White House officials communicated the decision to Republican leaders on Capitol Hill Monday, the people said.

The Justice Department, which was set to administer the fund, indicated Monday it would “abide by the Court’s ruling” and halt progress after a judge temporarily halted any payouts Friday. The statement otherwise defended the fund and did not say whether it could be revived later.

Earlier Monday, Trump met with Speaker Mike Johnson to discuss the controversial fund as part of a larger conversation about stalled legislation delivering some $70 billion in immigration enforcement funding for the remainder of Trump’s term.

During the meeting, Johnson raised concerns about the viability of the fund given the tenuous situation for the party-line funding bill, according to one of the people and two others.

Spokespeople for the White House and for Johnson did not respond to requests for comment Monday.

A planned Senate vote on the immigration bill was abandoned last month after the Justice Department announced the creation of the fund, which could be used to pay Trump allies who had been prosecuted under prior administrations. Democrats pledged to force votes on amendments related to the fund, creating a serious political dilemma for congressional Republicans.

Axios first reported on Trump’s decision Monday.

Two federal judges raised questions about the DOJ fund Friday, with one issuing a temporary restraining order preventing the Trump administration from disbursing any money from it. Separately, the judget who presided over the federal lawsuit filed by Trump that led to the fund’s creation launched an inquiry into the settlement deal.

Senate Majority Leader John Thune indicated the fund would have to be restricted in some way to unlock further progress on the GOP bill, saying in an interview it was a “safe bet” changes would be coming.

“The best way to handle it is if the administration decides to shut it down themselves,” he added in subsequent comments to reporters.

Alex Gangitano contributed to this report.

Senate Majority Leader John Thune said in an interview Monday it was a “safe bet” the GOP’s party-line immigration enforcement bill won’t move until Republicans get assurances on the new “Anti-Weaponization Fund” that could be used by the Justice Department to pay allegedly persecuted allies of President Donald Trump.

Thune said he is in talks with the Trump administration about securing those assurances.

“We’re in discussions and have been with them for a while now, so hope to have more on that soon,” Thune said when asked if he had gotten any indication the administration was going to propose changes to the $1.8 billion fund.

Senate GOP leaders delayed action on the immigration enforcement bill late last month after dozens of Republican members objected to the fund. Leaders believe several GOP senators would vote to include language in the bill that restricts the fund or blocks it entirely.

NEW YORK — Whether the hopefuls looking to replace retiring Rep. Jerry Nadler like it or not, their race has become a flashpoint for the national debate over how to regulate artificial intelligence — and Assemblymember Alex Bores has claimed the label of “AI guy.”

Much of Bores’ rise in the competitive campaign has stemmed from his signature policy in the state Legislature, the RAISE Act, one of the country’s landmark laws to establish guardrails for AI. It drew millions of dollars in attack ads against him from tech leaders who argue such regulations stymie innovation — and even more from those in the industry who are friendlier to AI oversight and are now boosting Bores.

Some of Bores’ opponents have sought to turn the tables on his AI message. In a recent interview, Kennedy scion Jack Schlossberg, one of Bores’ top opponents, criticized Bores’ RAISE Act as being “watered down” compared to how it was initially presented. He also went after the assemblymember’s promise to regulate the technology while being backed by those in the industry.

“When I hear someone talking about AI regulation who is being bankrolled by the AI industry, that reminds me of people who go in saying they’re gonna tackle fossil fuels and are actually taking money on the side from oil and gas companies,” Schlossberg said. “If we’re gonna be regulating AI, and we think it is the important issue that it is, we should have disinterested regulators, not those who have been bankrolled by one company and not another.”

Bores will often tout his record in the Legislature, where he has served since 2023. A common refrain is he was recognized by the Center for Effective Lawmaking as the “most effective new legislator” from the city and that he’s passed more than 30 bills.

But on what’s become his hallmark issue, Bores hasn’t always been a “yes.”

He has been supportive of dozens of AI-related bills, but on a handful of occasions, he’s been a detractor on bills aimed at addressing guardrails around technology and AI. That legislation includes:

• A4550, which would require the Department of Labor to study the long-term impact of artificial intelligence on the state workforce. Bores was the only one in the Labor Committee to vote no.
• A773, which would require some banks to conduct assessments on artificial intelligence in lending decisions and require applicants to consent to the technology. Bores was the only Democrat to vote against the legislation in the Codes Committee, joining Republicans.
• A3779, which would put restrictions on automated employment decision tools in hiring. He was the only no vote in both the Labor and Codes committees.
• A9430, also known as the LOADinG Act, which was touted as a “first in the nation legislation” to provide oversight of how state agencies use automated decision-making. Bores voted for it in the Science and Technology Committee, but then was one of two Democrats to vote against it on the floor. 
• A4983, which provides protections regarding health data. He voted for it in the Science and Technology Committee, and then against it in Codes. 

Bores expressed concerns about the specifics of these bills. With the bill on automated employment tools, for instance, he said the legislation “wasn’t addressing whether AI is the final decision, it was adding on a costly impact assessment without changing the actual standards of when bias occurs.” On the bank assessment bill, the assessments wouldn’t apply to national banks — and federal law already prohibits against discrimination for financial lending. A sticking point in the Labor Department bill was that AI use “to displace any human work” was prohibited until a study comes out at the latest in 2034.
“I think if we put forward requirements of assessments on small businesses that we’re not even estimating the cost of, then maybe we’re trying to talk about only a ban on small businesses while letting the biggest companies run wild,” he said. “You have to think about the impact of any legislation.”

Prior to his affirmative committee vote for the LOADinG Act, Bores raised concerns about applying too broad of a definition to biased outcomes, leading to his no vote on the floor. But he later approved the chapter amendment and A8295, spearheaded by the same sponsors of the LOADinG Act, which requires state government agencies to conduct impact assessments and disclose the use of automated decision-making tools.

Bores also had concerns about the definition of health data and a 24-hour waiting period to process health information in A4983. The next year, he voted in favor of a reworked version of the bill, which Gov. Kathy Hochul then vetoed for having “broad” definitions and scope, “creating potentially significant uncertainty about the information subject to regulation and compliance challenges for consumers, businesses and nonprofits.”

“AI is a deeply serious issue that requires legislators who are willing to dive into the details to make sure that New Yorkers and Americans are protected, and also that our laws are ones that actually achieve the goals they set out to do,” Bores said of his past votes. “I’ve always been a legislator that actually reads bills and actually picks up on specific revisions, and I engage with my colleagues in all of that, whether the topic is AI or anything else.”

Bores isn’t the only one who’s relying on his legislative record on tech and AI. A recent mailer from fellow Assemblymember Micah Lasher — another frontrunner for the seat — vowed the hopeful would “fight for a national ban on social media for kids under 16” and described him as a member of Congress who “will stand up to the tech titans.” The mailer also touts a long list of legislation he’s worked on that’s “protecting our kids from the dangers of social media and AI.” That includes a bill prohibiting AI chatbots from using features harmful to kids — legislation introduced by Bores that Lasher co-sponsored.

“When it comes to social media, it feels like we’ve been on autopilot for years,” Lasher writes in the mailer. “Letting a handful of tech giants make decisions — business decisions — that have enormous consequences for all of us, especially our kids.”

Schlossberg, who repeatedly goes after the AI-linked money pouring in to support Bores, has called for a federal investigation into rental car companies using AI to scan for damages and tack on fees. He also recently criticized Bores for missing AI-related votes in the Legislature, to which Bores replied, “I’m happy to put my legislative record up against anyone.”

A version of this report first appeared in New York Playbook. Subscribe here.