白宫与AI公司Anthropic举行会谈， amid政治紧张与五角大楼争端

White House meets AI firm Anthropic amid political tensions, Pentagon dispute

2026-04-17T16:10:50-04:00 / Fox News

Anthropic’s Mythos Preview model has uncovered thousands of previously unknown security flaws in critical systems

By Morgan Phillips Fox News

Published April 17, 2026 4:10pm EDT

Anthropic’s new AI model raises alarms over safety, cybersecurity concerns

OthersideAI co-founder and CEO Matt Shumer joins ‘The Sunday Briefing’ to discuss Anthropic’s new AI model, Mythos, and growing concerns over its advanced capabilities and potential cybersecurity risks.

NEW You can now listen to Fox News articles!

Listen to this article

8 min

One month after President Donald Trump ordered a government-wide halt on artificial intelligence firm Anthropic’s technology following a clash with the Pentagon, the company’s CEO is back at the White House for high-level talks — as officials reconsider whether a system they sidelined over national security and political concerns may be too important to ignore.

A source familiar with the meeting told Fox News White House chief of staff Susie Wiles met with Anthropic CEO Dario Amodei Friday.

Anthropic’s new artificial intelligence model, Mythos Preview, is considered so advanced that the company has restricted its release, limiting access to a small group of partners over concerns about potential misuse.

The meeting signals a rapid reversal inside the Trump administration, as officials weigh whether a system previously flagged as a national security risk could also be critical to defending U.S. infrastructure — exposing a growing internal tension over how to handle powerful AI tools with both defensive and offensive potential.

“Anthropic CEO Dario Amodei today met with senior administration officials for a productive discussion on how Anthropic and the U.S. government can work together on key shared priorities such as cybersecurity, America’s lead in the AI race, and AI safety. The meeting reflected Anthropic’s ongoing commitment to engaging with the U.S. government on the development of responsible AI. We are grateful for their time and are looking forward to continuing these discussions,” an Anthropic spokesperson told Fox News Digital.

MADURO RAID QUESTIONS TRIGGER PENTAGON REVIEW OF TOP AI FIRM AS POTENTIAL ‘SUPPLY CHAIN RISK’

The talks come despite a recent clash inside the Trump administration, as officials reconsider a company the Pentagon flagged as a supply chain risk. Its ties to former Biden officials and past criticism of Trump by its CEO have added a political dimension to the debate over whether its technology should return to government use.

A source familiar with the meeting told Fox News White House chief of staff Susie Wiles met with Anthropic CEO Dario Amodei Friday.(Chance Yeh/Getty Images for HubSpot))

That potential and the risks that come with it already have triggered tensions inside the U.S. government.

Pentagon clash, legal fight and reversal put Anthropic back in play

The meeting comes after a sharp break between Anthropic and the Pentagon earlier in 2026.

Defense Secretary Pete Hegseth designated the company a national security “supply chain risk,” effectively cutting it out of military systems and barring contractors from using its technology.

Anthropic is now challenging the designation in court, after filing multiple lawsuits against the Pentagon and other federal agencies arguing the “supply chain risk” label is unlawful and retaliatory.

The designation, which effectively bars contractors from using Anthropic’s technology and has been compared to measures typically reserved for foreign adversaries, already has faced conflicting rulings in federal court, with one judge temporarily blocking parts of the policy while an appeals court declined to halt its enforcement. The legal fight is ongoing, leaving contractors and agencies navigating uncertainty over whether and how Anthropic’s systems can be used.

The move followed a dispute over how the Pentagon could use Anthropic’s AI.

The company declined to grant open-ended authorization for “all lawful purposes,” instead insisting its systems not be used for mass domestic surveillance or fully autonomous weapons. While Pentagon officials said they do not rely on AI for either purpose, they rejected being constrained by a private company’s restrictions.

Trump then directed federal agencies to stop using Anthropic’s models altogether, escalating the standoff beyond the Defense Department into a government-wide halt.

Now, just weeks later, the company is back in high-level talks with the White House as officials weigh whether its new Mythos system — despite the earlier ban — could shift the balance of cyber defense and attack.

Political ties and past criticism may complicate White House talks

The dispute also has taken on a political dimension.

Amodei previously has drawn attention for his criticism of Trump, at one point likening him to a “feudal warlord” in a pre-2024-election Facebook post, according to a Wall Street Journal report.

In an internal message posted on Anthropic’s Slack platform and later leaked to The Information, Amodei suggested the Trump administration’s dispute with the company was driven in part by its refusal to offer what he described as “dictator-style praise.”

The message, written during a rapid escalation of tensions in early March, later was cited by the Wall Street Journal and other outlets. Amodei subsequently apologized for the tone, saying the post did not reflect his considered views.

FEDERAL APPEALS COURT REJECTS ANTHROPIC BID TO BLOCK PENTAGON BLACKLIST IN AI DISPUTE

When asked about Anthropic’s governance, hiring and broader political ties, a White House official said the administration “continues to proactively engage across government and industry to protect the United States and Americans,” including “working with frontier AI labs to ensure their models help secure critical software vulnerabilities.”

The official added that “any new technology that would potentially be used or deployed by the federal government requires a technical period of evaluation for fidelity and security,” and said “the collective effort of all involved will ultimately benefit industry, and our country, as a whole.”

Amodei previously has drawn attention for his criticism of Trump, at one point likening him to a “feudal warlord” in a pre-2024-election Facebook post, according to a Wall Street Journal report.(Patrick Sison/AP Photo)

Beyond the immediate dispute, the company’s broader ties to Washington also have drawn attention.

Anthropic’s governance structure has also drawn attention as the administration weighs closer engagement. The company is overseen in part by an independent “Long-Term Benefit Trust,” an unusual mechanism designed to give nonfinancial stakeholders influence over corporate decisions.

The trust holds special voting shares that allow it to appoint and eventually control a majority of the company’s board, with members drawn from national security, public policy and global development backgrounds.

Current trustees include Clinton Health Access Initiative CEO Neil Buddy Shah, Carnegie Endowment president Mariano-Florentino Cuéllar, a Democrat who was appointed to the California Supreme Court by former Gov. Jerry Brown in 2014, and Center for a New American Security CEO Richard Fontaine — who advised John McCain’s 2008 presidential campaign. The group is a mix of policy and national security leaders that underscores the company’s deep ties to Washington and global policy circles.

Anthropic’s backers also have placed it at the center of overlapping tech, policy and political networks.

Early funding for the company included investments from figures such as Facebook co-founder Dustin Moskovitz and former Google CEO Eric Schmidt, both longtime Democratic donors, and a major early investment from Sam Bankman-Fried’s FTX.

At the same time, the company has since attracted a broad range of major institutional investors — including Amazon, Google and Microsoft — reflecting its growing role in the global AI race and complicating efforts to characterize it along purely political lines.

The company also has brought on several officials from the Biden administration into key policy roles, further embedding Anthropic in Washington’s AI policy ecosystem. Among them is Tarun Chhabra, a former National Security Council official who now leads the company’s national security policy work, as well as other advisers and staff with experience shaping federal AI and technology strategy.

Anthropic also has sought to build ties across party lines as it expands its presence in Washington.

The company employs policy staff with Republican backgrounds, including legislative analyst Benjamin Merkel and lobbyist Mary Croghan, and in February added Chris Liddell — a former deputy White House chief of staff under Trump — to its board. It has contributed $20 million to Public First Action, a bipartisan group that backs candidates from both parties who support AI regulation.

A federal judge’s decision to block the Trump administration from banning AI firm Anthropic from Department of War use is igniting a debate over whether the ruling pushes courts into national security decision-making.(Samyukta Lakshmi/Bloomberg via Getty Images; Eugene Hoshiko/Pool/Reuters)

The company has also faced criticism from within the Trump administration.

White House AI adviser David Sacks has accused Anthropic of pursuing a “regulatory capture” strategy, arguing the firm is using concerns about AI safety to push rules that could benefit its own position while slowing competitors.

Anthropic has pushed back on those claims, saying its approach reflects genuine concerns about the risks posed by advanced AI systems.

JUDGE FREEZES TRUMP ADMIN MOVE AGAINST AI FIRM, FUELING BATTLE OVER SECURITY AUTHORITY

New AI system could reshape cyber warfare, raising alarms inside US government

The new technology could help developers identify and fix long-standing security flaws, but it could also give hackers a powerful new tool to target U.S. businesses and government systems.

“Given the rate of AI progress, it will not be long before such capabilities proliferate, potentially beyond actors who are committed to deploying them safely,” Anthropic said in its announcement. “The fallout — for economies, public safety, and national security — could be severe.”

Anthropic has not released Mythos publicly, instead limiting access through a program called Project Glasswing, where a select group of companies use the model to scan critical systems for vulnerabilities.

Pages from the Anthropic website and the company’s logos are displayed on a computer screen in New York on Thursday, Feb. 26, 2026.(Patrick Sison/AP Photo)

CLICK HERE TO DOWNLOAD THE FOX NEWS APP

The company says the system has already uncovered thousands of previously unknown flaws — some decades old — underscoring both its defensive value and the risk it could be used to accelerate cyberattacks if the technology spreads.

Fox Business’ Edward Lawrence contributed to this report.