If you picked the ChatGPT leak, you understand corporate espionage. The answer is the first option. Grok manufactured Category A illegal material. AI stopped passively hosting bad actors. The models actively created the contraband. A technicality about human interaction could no longer shield the platform owners. They were dragged into the spotlight. Regulators confronted Elon Musk's X in January 2026 over Grok's output. The company deployed a rapid safeguard.