Cryptographer Matthew Green warns that AI could accelerate the discovery of vulnerabilities in public-key encryption faster than human standards bodies can respond, estimating a 15% chance of functionally losing confidence in existing algorithms. He references Impagliazzo's 'Minicrypt' scenario where public-key encryption becomes impossible, emphasizing that preparation must happen in advance since AI's pace of producing surprises far outstrips human institutional response times.
- Source
- Simon Willison
- Published
- Oct 9, 2026 at 11:02 PM
- Score
- 7.0 / 10