The Mathematical AI Safety Institute wants to prove AI is safe the way cryptographers prove codes are unbreakable

Fields Medal recipient Jacob Tsimerman has established the Mathematical A.I. Safety Institute to develop formal proofs for the security of artificial intelligence systems. By applying the rigorous methodologies used in cryptography, the institute aims to create mathematical guarantees that AI models will behave predictably and securely. This approach shifts the focus of safety from empirical testing to verifiable, objective standards, potentially addressing long-standing concerns regarding the reliability of complex machine learning architectures.