Home / Jul 29, 2026 / Story
0
#7 The Hacker News general July 28, 2026 at 18:59 UTC

Claude AI Just Cracked a Post-Quantum Test Scheme and Found a Faster 7-Round AES Attack

By [email protected] (The Hacker News)

AI Summary

Anthropic's Claude Mythos Preview AI model derived an end-to-end key-recovery attack against HAWK-256, a post-quantum signature scheme candidate, exploiting a previously unused symmetry in its underlying lattice structure, with a full attack runtime of approximately 3 hours 42 minutes on a 96-core server. The model also achieved a 200x–800x speedup for an existing attack against 7-round AES-128. This marks the first publicly documented case of a frontier AI model discovering novel cryptanalytic attacks against modern cryptographic primitives.

Relevance score: 84.0/100

# More from July 29