The Indicator from Planet Money
The Indicator from Planet Money

The new AI model that could steal your life savings

May 11, 2026

AI Summary

5 min read

In April, Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell called a special meeting with bank CEOs. They wanted to make sure banks were keeping their digital vaults secure in an age of rapidly advancing AI. The trigger was an announcement from the AI company Anthropic: it had developed a new model so state-of-the-art, so clever, so dangerous that it couldn't yet release it to the wider public. Anthropic called the model Claude Mythos.

The model that chains vulnerabilities

Claude Mythos is, in some ways, a typical language model. It wasn't fundamentally different from a chatbot that can answer general questions or write lines of code, like Anthropic's other Claude models or ChatGPT. It hadn't been trained specifically for cybersecurity. It could write poetry, limericks, or give you a new recipe for chocolate chip cookies. But when Anthropic was testing Mythos, people inside the company like Michael Moore, who heads up the company's cybersecurity products, started getting concerned. "It showed a really advanced capability of chaining together individual vulnerabilities into an exploit that was potentially really dangerous," Moore says.

Continue reading the full summary in the app — free to try.

Read Full Summary →

Free • No credit card required

Listen to Audio Summary Open in App

What you'll learn

  • 1 (00:00) **Treasury and Fed Call Emergency Meeting Over AI** - Secretary Bessent and Chair Powell convene bank CEOs to discuss securing digital vaults against advanced AI threats.
  • 2 (01:07) **The Core Question: Is Our Money Still Safe?** - Hosts frame the episode around whether bank accounts remain secure in the age of increasingly capable AI.
  • 3 (02:32) **What Makes Mythos Different?** - The model is a typical language model at base, but its ability to chain vulnerabilities together sets it apart.
  • 4 (03:48) **Project Glass Wing: Giving Banks a Head Start** - Anthropic grants early access to Mythos only to a small group of financial institutions to patch vulnerabilities before wider release.
  • 5 (04:17) **Should People Worry About Their Life Savings?** - The host presses Moore on whether the limited early access is sufficient to protect the public.
  • 6 (05:06) **The Logistical Challenge of Warning the World** - The hosts question how Anthropic can possibly warn all software users before Mythos gets into the wrong hands.
  • 7 (05:56) **Anthropic's New Product: Claude Security** - The company releases a special tool for finding and fixing software vulnerabilities, though not yet based on the cutting-edge Mythos model.

+ Full timestamped outline available in the app

Guests on this episode

Michael Moore

Show Notes

Anthropic’s AI model, Claude Mythos, is very powerful. SO powerful it can find software vulnerabilities that might let it, I don’t know, steal your bank login information. Anthropic is holding back the model from a wider release for now. Today on the show, how worried should we really be about Mythos? And are its capabilities actually unique?

The Indicator has a weekly newsletter! Sign up now: npr.org/indicatornewsletter 

Related episodes:  
Fighting AI with AI 
How AI might mess with financial models

For sponsor-free episodes of The Indicator from Planet Money, subscribe to Planet Money+ via Apple Podcasts or at plus.npr.org. Fact-checking by Sierra Juarez. Music by Drop Electric. Find us: TikTok, Instagram, Facebook, Newsletter.  


See pcm.adswizz.com for information about our collection and use of personal data for sponsorship and to manage your podcast sponsorship preferences.

NPR Privacy Policy
The Indicator from Planet Money

More from this podcast

The Indicator from Planet Money →