Legal frameworks in the US and elsewhere continue to treat large language models and other AI systems as tools rather than independent entities, directing criminal liability toward developers, operators, or users instead. Recent state legislation explicitly denies AI legal personhood, while probes such as Florida’s investigation into OpenAI focus on company responsibility for chatbot outputs rather than charging the model itself. This structure underpins the 94.9% market-implied probability against charges before 2027. A realistic challenge would require either an unprecedented grant of limited personhood or a high-profile autonomous incident that forces rapid statutory change, both of which lack precedent or momentum in current regulatory trajectories.