Cryptographer warns AI could accelerate collapse of public-key encryption standards

Cryptographer Matthew Green raised concerns about the potential for AI to destabilize public-key encryption standards, estimating a 15% chance that society could lose confidence in existing algorithms. In a post on Simon Willison’s Weblog, Green argued that the speed of AI-driven surprises far outpaces the ability of human institutions to replace or update cryptographic standards, even with AI assistance.

Green also referenced a 1% probability of inhabiting 'Minicrypt,' a hypothetical world proposed by Russell Impagliazzo in which public-key encryption is fundamentally impossible. He framed these risks as worst-case scenarios but stressed the need for proactive preparation to mitigate potential fallout.

The remarks were originally shared on Twitter and later quoted by Simon Willison, who added context about Minicrypt and its implications for cryptographic security. Green’s comments underscore broader anxieties about AI’s role in accelerating unforeseen technological disruptions that could outstrip the capacity for institutional adaptation.