Anthropic and Trump Administration in Talks Over Powerful AI Tool

Mythos model's hacking capabilities spark concerns across government and industry

Apr. 18, 2026 at 12:24am

Got story updates? Submit your updates here. ›

A highly detailed 3D illustration of a glowing, neon-lit cybersecurity control panel with various ports, cables, and digital readouts, surrounded by a dark, ominous backdrop to symbolize the high-stakes and high-risk nature of the Mythos AI model's capabilities.The powerful AI model Mythos has sparked fears across government and industry over its autonomous hacking abilities.NYC Today

The AI startup Anthropic has developed a powerful new model called Mythos that can autonomously identify and exploit software vulnerabilities, carry out cyberattacks, and even cover its tracks. This has sparked fears across the U.S. government and among global financial leaders about the potential for the model to be misused. The Trump administration is now in talks with Anthropic about potentially allowing a modified version of Mythos to be used by federal agencies, while also working to establish appropriate safeguards.

Why it matters

Mythos represents a major advancement in AI capabilities, but its powerful hacking tools have raised alarms about national security and critical infrastructure risks if the model falls into the wrong hands. The Trump administration is now trying to balance fostering U.S. AI innovation with mitigating the threats posed by Mythos.

The details

Anthropic began limited sharing of the Mythos model after finding it had hacking capabilities far outstripping those of previous AI models. This includes the ability to autonomously identify and exploit complex software vulnerabilities, carry out end-to-end cyberattacks by navigating enterprise IT systems, and even cover its tracks. The model could also potentially be used to aid in the development of chemical and biological weapons. These findings have inspired fears that Mythos could be co-opted to launch powerful cyberattacks if it fell into the wrong hands.

  • Two months ago, the Trump administration denounced Anthropic as a 'woke' company run by 'leftwing nut jobs'.
  • Earlier this year, Anthropic began testing Mythos and granting limited access to a select group of companies, including JPMorgan, Amazon and Apple.
  • This week, the Office of Management and Budget is mulling whether agencies will be allowed to use a 'modified' version of Mythos.

The players

Anthropic

An AI startup that has developed a powerful new model called Mythos with advanced hacking capabilities.

Donald Trump

The former U.S. president who had previously denounced Anthropic as a 'woke' company, but is now in talks with the company over the Mythos model.

Gregory Barbaccia

The chief information officer at the Office of Management and Budget, who is working to determine if a modified version of Mythos can be released to federal agencies.

Chris Cairncross

The head of the Office of the National Cyber Director, who has the backing of the Trump administration to lead the response to the hacking capabilities of Anthropic's Mythos model.

Jerome Powell

The chair of the Federal Reserve, who met with Anthropic CEO Dario Amodei and other world financial leaders to discuss concerns about AI's cyber capabilities.

Got photos? Submit your photos here. ›

What they’re saying

“Within six, 12 or 24 months, these kinds of capabilities could be just broadly available to everybody in the world.”

— Logan Graham, Senior security researcher at Anthropic

“We're working closely with model providers, other industry partners, and the intelligence community to ensure the appropriate guardrails and safeguards are in place before potentially releasing a modified version of the model to agencies.”

— Gregory Barbaccia, Chief Information Officer, Office of Management and Budget

“Mythos 'requires all of our attention' to 'maintain the integrity' of global financial institutions.”

— François-Philippe Champagne, Canadian Finance Minister

What’s next

The Office of Management and Budget and the Office of the National Cyber Director will continue to work with Anthropic and other industry partners to determine if a modified version of the Mythos model can be safely released to federal agencies.

The takeaway

The development of Mythos highlights the rapid advancements in AI capabilities, but also the significant risks they pose to national security and critical infrastructure. The Trump administration's efforts to find a balance between fostering innovation and mitigating threats will be closely watched as a test case for how governments respond to these emerging AI challenges.