← Home
The Indicator from Planet Money · May 11, 2026 · 9m

The new AI model that could steal your life savings

Show Notes

Anthropic has developed an AI model called Claude Mythos that is powerful enough to find software vulnerabilities, enabling it to potentially steal bank login information. The company is holding back the model from a wider release for now. This episode from NPR's The Indicator from Planet Money examines how worried we should be about Mythos and whether its capabilities are actually unique among current AI models.

This summary was generated from show notes and public descriptions, not from a full transcript review. Details may contain inaccuracies.

Highlights

•
Claude Mythos finds security vulnerabilities
Anthropic's Claude Mythos AI model demonstrates the ability to find software vulnerabilities that could be used to steal bank login information.
ClaudeChatGPTGemini
•
Anthropic withholds Mythos from wider release
Anthropic is holding back Claude Mythos from a wider release due to its potentially dangerous capabilities.
ClaudeChatGPTGemini
•
How unique is Mythos's capability?
The episode explores whether Claude Mythos's vulnerability-finding powers are unique or part of a wider trend among AI models.
ClaudeChatGPTGemini
Was this useful?