June 26, 2026

Anthropic’s Mythos model raises alarm for Rep. Andrew Garbarino over AI threats to bank security

Rep. Andrew Garbarino (R-N.Y.) expressed significant concern after a closed-door demonstration of Anthropic’s new AI model, Mythos, which showcased its capability to exploit vulnerabilities in banking systems. Garbarino reported that the AI was able to empty bank accounts and then identify and fix the vulnerabilities it exploited. He highlighted the urgent need for government action in response to the threats posed by such powerful AI technologies. This aligns with President Trump’s executive order aimed at ensuring federal agencies can access advanced AI models before their public release, as well as ongoing bipartisan efforts in Congress to develop a regulatory framework governing AI advancements.

Anthropic: Anthropic is an AI company focused on developing advanced frontier models with an emphasis on safety and alignment. Its latest Claude Mythos series demonstrates exceptional capabilities in cybersecurity tasks such as autonomously identifying software vulnerabilities and generating exploits. In the news, Anthropic conducted a closed-door demonstration of Mythos for congressional members including Rep. Garbarino, illustrating the model’s potential to execute cyber operations like emptying bank accounts and prompting calls for greater government oversight before public release.
Lori Trahan: Lori Trahan is a Democratic U.S. Representative from Massachusetts involved in technology and consumer protection issues. She has worked on bipartisan efforts to address AI governance and safety. In the news, she is cited alongside Rep. Obernolte for a recent discussion draft proposing a federal framework for AI regulation.
Donald Trump: Donald Trump is the current President of the United States. He recently issued an executive order directing federal agencies to prepare for AI-related cyber threats and establishing a voluntary framework for AI companies to share model access with the government. The news notes Garbarino’s praise for this order as a step toward managing emerging AI risks.
Jay Obernolte: Jay Obernolte is a Republican U.S. Representative from California focused on technology and innovation policy. He has collaborated on legislative proposals to create structured oversight for AI development and deployment. In the news, Garbarino referenced Obernolte’s recent joint discussion draft with Rep. Trahan aimed at establishing a federal framework for AI.
Andrew Garbarino: Andrew Garbarino is a Republican U.S. Representative from New York who chairs the House Homeland Security Committee. He has publicly highlighted risks from advanced AI after participating in demonstrations showing models’ ability to bypass safeguards and perform autonomous actions. In the news, Garbarino described Anthropic’s Mythos demo and urged swift government action on AI threats while acknowledging uncertainty about specific regulatory solutions.

`json
{
“Policy”: “President Trump’s executive order sets up mechanisms for federal agencies to access advanced AI models prior to broader release in order to address potential cyber risks.”,
“Regulation”: “Bipartisan congressional discussions are advancing draft legislation to create a national framework governing the development and deployment of frontier AI systems.”,
“Cybersecurity”: “Anthropic has limited public access to its Mythos model and is testing it through controlled initiatives with select organizations to explore defensive applications against cyber threats.”
}
`

AI tools built by Emerald Force

Built and supported by Emerald Force.

You might also like