Google DeepMind: AI Discovers Two Weaknesses in Legacy 1024-bit RSA Systems

TechFlow, November 7: Google DeepMind disclosed that, in its research, it used human-guided AI agents to discover two significant weaknesses in legacy 1024-bit RSA cryptographic systems, reducing the computational cost required to break them below previous expectations.

Pushmeet Kohli, Vice President of Science at Google DeepMind, said in the research briefing that one finding involves digital signature forgery; another new algorithm reduces the average cost of each additional key when factoring keys of the same bit length in bulk. The research affects only deprecated 1024-bit RSA systems; modern cryptographic protocols using 2048-bit or longer keys remain secure.

Source: TechFlow