SHIFTHEADS: The AI Upgrade You Lost Without Warning

Something changed in your Claude window this week and technology analyst Carmi Levy is here to tell you it was not an accident. It was a decision, made by a government that is not yours, about a tool you depend on.   The AI you were using had already been dialed back before it reached you. The full-strength version was quietly restricted to a handful of trusted governments and researchers under something called Project Glasswing because it was capable of breaking through existing cybersecurity protections. What you had was the safer version. Then the US government took that too. Carmi Levy lays out what is technically at stake, then puts the politics on the table, including whether this is really about cybersecurity or about which tech companies have been playing nice with Washington.   Canada has a seat at this table. Whether Ottawa uses it is the question worth asking right now.   Topics: Claude AI, Anthropic, Project Glasswing, AI cybersecurity, US AI policy   GUEST: Carmi Levy   Originally aired on 2026-06-16