Aaronson: AI could break encryption within two years

Former OpenAI safety researcher and quantum computing scholar Scott Aaronson said AI models may crack one or more widely used encryption algorithms within one to two years. He argued OpenAI deliberately avoided cryptography among the nearly 400 math problems its models solved.

• Aaronson estimates AI could break common ciphers in one to two years
• OpenAI published about 400 AI-solved math problems with no cryptography
• Breaking RSA or Bitcoin encryption would require warning the whole world, not one company
• AI has already produced a proof of the Unique Games Conjecture studied since 2009