Cryptographers debate whether AI math advances threaten public-key cryptography
Original title: Do AI’s mathematical advances put PKI under threat?
Matthew Green puts a 15% chance on losing confidence in current public-key algorithms; peers say follow the post-quantum transition.
Johns Hopkins cryptographer Matthew Green warned that machines outperforming humans on some math problems could erode confidence in standardized public-key encryption and some signatures that underpin PKI, TLS, and code signing. He estimated a 15% chance of functionally losing that confidence and called for cryptographic agility. Markku-Juhani O. Saarinen said AI helped break HAWK and shorten McEliece parameters in 2026, but advised against hasty replacement. Alan Woodward said large language models have not yet deeply dented cryptography.
Unconfirmed: Green’s probability estimates are his assessment; the article reports no confirmed break of standardized public-key algorithms.
- Green estimated a 1% chance of living in Minicrypt.
- Germany’s cyber agency withdrew its Classic McEliece recommendation at the end of September.
- Adjusting HAWK and McEliece parameters would roughly double keys, signatures, and ciphertexts.
- Saarinen said RSA and elliptic-curve systems have known polynomial-time quantum attacks since 1994.
- ML-KEM, ML-DSA, and SLH-DSA have had more than a decade of analysis.
- 1.Follow existing post-quantum transition standards rather than hasty cryptographic replacements.
- Organizations
- Johns HopkinsTampere UniversityUniversity of Surrey
From the original article
5 min read at csoonline.com
Warnings by a leading cryptographer that industry-wide public-key cryptography could fail because of advances in AI has sparked a debate in the security community. In a series of post on X, Johns Hopkins cryptographer Matthew Green warned that existing public-key cryptography systems were at risk because machines have begun outperforming humans in their ability to address some mathematical problems.
csoonline.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.