Claude identifies cryptographic flaws in HAWK and AES variants
Anthropic researchers leveraged Claude to identify mathematical vulnerabilities in established cryptographic systems, including weaknesses in HAWK and a reduced-round variant of AES. While the findings carry no immediate security implications for deployed systems, the work demonstrates LLMs' emerging capacity in formal mathematical reasoning and cryptanalysis. The research surfaces a critical methodological insight: prompting strategies matter significantly when directing models toward problems they initially perceive as unsolvable, suggesting that capability often hinges on framing rather than raw model limitations. This positions Claude as a tool for security researchers and hints at broader applications in mathematical discovery.
Modelwire context
ExplainerThe most underreported detail is the prompting methodology itself: researchers found that models would abandon cryptanalysis tasks when framing suggested the problem was intractable, but succeeded when framing implied solvability was expected. That's not a footnote about UX, it's a substantive finding about how capability assessments can systematically undercount what models can do.
This is largely disconnected from recent activity in our archive, as we have no prior coverage to anchor it to. It belongs to a broader conversation happening across academic and security research communities about whether LLMs can contribute meaningfully to formal mathematics, a space where skepticism has historically been well-founded given models' tendency to confabulate proofs. The reduced-round AES finding is the kind of bounded, verifiable result that gives that conversation more traction than benchmark claims typically do, because the output can be checked independently.
Watch whether independent cryptographers attempt to reproduce or extend these findings on full-round primitives in the next six months. Replication on a non-reduced variant would substantially change how the security research community treats LLMs as analytical tools.
This analysis is generated by Modelwire’s editorial layer from our archive and the summary above. It is not a substitute for the original reporting. How we write it.
MentionsAnthropic · Claude · Simon Willison · HAWK · AES · Claude Mythos
Modelwire Editorial
This synthesis and analysis was prepared by the Modelwire editorial team. We use advanced language models to read, ground, and connect the day’s most significant AI developments, providing original strategic context that helps practitioners and leaders stay ahead of the frontier.
Modelwire summarizes, we don’t republish. Simon Willison originally reported this story as “Discovering cryptographic weaknesses with Claude”. The full content lives on simonwillison.net. If you’re a publisher and want a different summarization policy for your work, see our takedown page.