Source: Washington Post
by James Harkin
“What recourse do people have when encryption technology that they thought was pretty good comes up against an AI program capable of impressive mathematical reasoning? In a July blog post, Anthropic disclosed that its researchers, using their Claude Mythos Preview model, had discovered improved ways to attack cryptographic algorithms (the mathematical methods used to keep online data private) — a development with transformative implications for cryptography. The concern here isn’t just about protecting data — it’s also about protecting identities. In the last few decades, society has found sophisticated ways to blur or redact things to protect people’s anonymity. But as it begins to flex its muscles, AI may be able to de-blur them instantly.” (09/08/26)