'Is your AI agent Pomeranian or wolf?' asks Harvard expert in test for accountability

Harvard professor Jordi Weinstock has proposed a "Canine Agentic Framework" to determine accountability if an AI agent causes harm. Based on how "domesticated" or "dangerous" an AI agent is, it classifies them as a Pomeranian, pitbull, fox or wolf. "If a Pomeranian bites you...you sue its owner. But if a wolf bites you, there's no one to sue," he said.