[Google DeepMind: AI Discovers Two Weaknesses in Older 1024-bit RSA System] According to TechFlow, on November 7th, Google DeepMind disclosed that its research using human-guided AI agents has discovered two major weaknesses in the older 1024-bit RSA cryptosystem, making the computational cost of breaking it lower than previously expected. Pushmeet Kohli, Vice President of Science at Google DeepMind, stated in the research description that one finding relates to digital signature forgery; the other new algorithm reduces the average cost per additional key when batch-factoring keys of equal length. The research only affects the obsolete 1024-bit RSA system; modern encryption protocols using 2048-bit or longer keys remain secure.