Cryptographer slams Anthropic for dismissing replay-attack encryption bugs he reported in May

Apoorva__Lal · x · 2026-09-12

Cryptographer Matthew Green publicly called out Anthropic, saying he reported encryption issues exploitable via replay attacks in May only to be told they were outside the company's threat model — and that people apparently exploited them for months afterward. Matt Parlmer's viral repost argued the lapse looks like gross negligence, or even an attempt to invite maximal government regulation of the sector after commercial wins. The incident puts a spotlight on how AI labs handle external security reports.

Related event: Cryptographer Slams Anthropic for Ignoring Reported Flaw(2 posts)→

Original post →

More from Companies & People

Companies & People channel →