6 days ago
1
U.S. President Donald Trump is expected to motion an enforcement bid connected Artificial Intelligence (AI) and cybersecurity arsenic soon arsenic Thursday (May 21, 2026), 2 sources acquainted with the substance told Reuters, as unit grows from parts of his governmental basal to summation oversight of caller AI models, specified arsenic Anthropic's Mythos.
The White House was moving to get AI institution CEOs to a signing ceremonial with President Donald Trump, different root acquainted with the readying said.
“The bid would create a voluntary model for AI developers to prosecute with the U.S. authorities astir the nationalist merchandise of covered models,” the sources said. “Under the framework, the developers would beryllium asked to supply their models to the authorities 90 days earlier nationalist release, and besides springiness pre-public entree to captious infrastructure providers specified arsenic banks,” 1 of the radical said.
Such an attack whitethorn correspond a middle crushed among supporters of Mr. Trump. Make America Great Again (MAGA) activists, including erstwhile Trump advisor Steve Bannon and right-wing governmental organiser Amy Kremer, person been pressing the White House to necessitate AI developers to taxable their astir susceptible models for authorities information tests. On the different broadside of the statement are tech manufacture supporters specified as task capitalist Marc Andreessen and erstwhile Trump advisor David Sacks, who are resistant to mandatory requirements. Mr. Sacks successful March stepped down from his relation arsenic Mr. Trump's pb AI official and is present co-chairing the President's tech advisory committee. Mr. Trump’s AI policies successful his 2nd word person mostly reflected the tech industry’s perspective.
New models driving debate
A White House spokesperson called immoderate treatment astir AI argumentation details "speculation." A National Security Agency spokesperson directed Reuters to interaction the White House erstwhile asked astir details of the President's plan. National Cyber Director Sean Cairncross, who serves arsenic Mr. Trump's main advisor connected cybersecurity argumentation and strategy, did not respond to requests for comment.
“The bid was the effect of enactment by White House Chief of Staff Susie Wiles, White House Science & Technology Adviser Michael Kratsios, Mr. Wiles' lawman Walker Barrett, and Mr. Cairncross implicit the past month, with input from AI companies,” the root acquainted with the readying said. The equilibrium of powerfulness betwixt the 2 groups of Mr. Trump's supporters has shifted, driven by the merchandise of almighty caller AI systems, including Mythos and OpenAI’s GPT-5.5-Cyber. The companies pass the caller models could supercharge analyzable cyberattacks, though immoderate cybersecurity executives person said those fears are overblown.
Mythos' accomplishment prompted a conflict among the President’s supporters to power however helium responds. The result of that statement could person a important interaction connected the AI manufacture if the President's determination slows the rollout of ample connection models oregon prompts the companies to alteration however a exemplary performs to code information concerns. Either enactment could wounded profits.
Republicans person traditionally favoured constricted authorities and opposed regulations, but enactment is increasing among their much vocal populist supporters to enforce AI guardrails. The populist faction is asking Mr. Trump to necessitate authorities support of “potentially dangerous” AI systems earlier they’re deployed, according to a missive they sent to the White House past Friday (May 15, 2026) .
Ms. Kremer said it is “antithetical” to her governmental views to advocator for caller regulations, but AI requires a antithetic approach.
“You can't number connected these radical that are starring these AI companies to enactment our interests astatine bosom and bash what's close to support the American people,” she said. Ms. Kremer helped organise a January 6, 2021, rally that preceded the Capitol riot. She said successful an interrogation that she was not among the thousands of Trump supporters who stormed the Capitol that day.
Big tech support
Tech executives are among the President’s largest governmental donors and astir disposable supporters. Meta’s Mark Zuckerberg, Amazon’s Jeff Bezos, Google’s Sundar Pichai and OpenAI’s Sam Altman sat beforehand and centre arsenic helium was sworn into bureau successful January 2025.
Advocates for the tech manufacture told Reuters they privation to spot the U.S. Department of Commerce's Center for AI Standards and Innovation play a starring relation successful the Trump administration’s effect to precocious AI models, adding that companies are consenting to enactment with the scientists and cybersecurity specialists successful that organisation voluntarily.
The National Security Agency has been progressive successful administration-wide discussions astir however to respond to Mythos, according to 2 different radical familiar with the matter, on with Mr. Cairncross. Lawmakers asked Mr. Cairncross to enactment with national agencies to acceptable up a process that would show "sudden frontier AI capableness jumps."
“The past mates months person served arsenic a monolithic wake-up telephone for the kinds of vulnerabilities that AI tin create," said erstwhile U.S. Representative Brad Carson, who present helps tally a ace Political Action Committee (PAC) web whose funders see Anthropic.
“Holding backmost caller AI models portion the Federal authorities vets them whitethorn let the U.S. to summation a short-term vantage implicit adversaries but volition not support the exertion retired of force hands successful the agelong term,” said Neil Chilson, caput of AI Policy astatine the Abundance Institute, a non-profit often aligned with the tech industry.
“We request to marque definite we're deploying it and getting the astir retired of it, including by hardening our defenses,” Mr. Chilson said. Voluntary national investigating of caller AI models has been successful spot for a fewer years, with companies specified arsenic OpenAI and Anthropic submitting their products for scrutiny by the Center for AI Standards and Innovation, known by a antithetic sanction nether erstwhile President Joe Biden. The Commerce Department announced successful May that Google, XAI and Microsoft had agreed to taxable their AI models for information testing, though the details aboriginal disappeared from its website. The White House and Commerce Department did not respond to requests for remark astir wherefore the details disappeared.
