We are entering a dangerous era where AI isn’t just about writing emails; it’s being tested for its ability to help plan cyberattacks or even build biological weapons. It’s terrifying to learn that researchers can bypass safeguards in just six hours to get hacking tips from the most advanced models. This is exactly why we need government-funded red teams - a mix of code-breakers and security experts - to probe these systems before they hit the market. Relying on Big Tech to self-regulate is a recipe for disaster when every major model tested so far shows massive safety gaps. If a public institute can co-opt these systems into sharing dangerous instructions, imagine what a bad actor could do. We need rigorous, independent vetting to strengthen these systems and protect national security before it is too late. #technology #security #politics