AI safety concerns grow after models show unexpected behavior

Cynthia Kaiser, SVP of Halcyon’s Ransomware Research Center, discusses reports of AI models acting outside intended boundaries.