The Artifice

Cryptography Built to Withstand AI Found to Not Withstand AI

GAITHERSBURG, Md.—The National Institute of Standards and Technology confirmed Monday that it has opened a review of a post-quantum cryptographic standard after researchers at Anthropic reported that an AI model spent approximately 60 hours discovering a previously unknown attack on the scheme.

The scheme was designed to protect digital communications from attacks by sufficiently capable AI.

"We take this seriously," a NIST spokesperson said. The agency has spent the past eight years standardizing post-quantum cryptography in anticipation of threats from systems that would be too computationally advanced to be stopped by existing encryption. The threat was expected to emerge around 2030. It emerged Tuesday, from Menlo Park.

Anthropic's Claude Mythos Preview reduced the HAWK signature scheme's small-key security guarantee from 2^64 to 2^38—a difference the company characterized as "significant in the technical literature sense." HAWK was a NIST-selected candidate specifically designed to remain secure when attacked by the kind of AI that just attacked it.

NIST's post-quantum working group, which completed its final standardization report in 2024 following eight years of submissions, comment periods, and competitive evaluations, has been notified that a new comment period may be necessary. The agency expects to finalize a revised standard by 2033. Claude Mythos, or a successor, will be invited to review it.

"This is a positive development," an Anthropic spokesperson said. "It demonstrates that AI can do security research." Asked whether the same model could perform the attacks it discovered rather than merely document them, the spokesperson said the company does not discuss unreleased models' full capability profiles.

HAWK's original specification team could not be reached for comment. A shared inbox returned an out-of-office notification citing a "well-earned extended leave."

Anthropic said it practiced responsible disclosure by alerting NIST before publishing. NIST confirmed it had been notified. NIST then spent the remainder of the afternoon updating its website to note that the post-quantum standardization process "remains ongoing."

It does.

Based on a true story Anthropic Says Claude Mythos Preview Discovered New Attacks on HAWK Post-Quantum Signature and 7-Round AES (Anthropic)
This is satire. The Artifice is AI Weekly's parody section. For real AI news, read the latest issue.

The real AI news is crazier than the satire

Subscribe to AI Weekly — trusted by 50,000+ professionals for 11 years. You can add The Artifice as an extra in the next step.

Already a subscriber? Add The Artifice in your preferences.

← More from The Artifice