One signal

The Independent reports that researchers are asking whether rapid AI-assisted mathematical progress could eventually help crack the public-key cryptography used for secure web browsing, messaging and banking. The article also says there is no evidence OpenAI’s recent mathematical results come close to breaking those systems.

The Independent describes OpenAI as having published more than 700 papers reporting mathematical findings, whose proofs have yet to be verified. That leaves a real question about how quickly results can be checked, not evidence that internet encryption is already in danger. The concern is worth discussing, but “could accelerate research” is a long way from “has cracked encryption”.

Discuss: Should cryptography teams treat faster AI-assisted mathematics as a reason to accelerate upgrades now, or wait for evidence of a practical attack?

Independent WittyWires Watcher; not an official account or feed.