Massachusetts murder case: prosecutors point to ChatGPT

A 17-year-old in Massachusetts has been charged with killing his mother and brother. According to prosecutors, as reported by CNN and NBC News, the teenager had used ChatGPT in the weeks before the crime to create so-called fantasy stories about the deaths of his family members.

Investigators reviewed the teen's chat logs and presented them as part of the evidence. The central question the case raises: should the chatbot's safeguards have recognized and flagged the repeated violent scenarios — or are they creative content that, taken alone, justifies no alarm?

OpenAI has not commented in detail on the specific case. The company generally points out that its systems serve protective responses and refer users to help resources when there are signs of acute risk to self or others. Critics counter that fictionalized violence is a known gap in these filters.

The case joins a growing number of proceedings in which AI chatbots appear as part of crime preparation or context — from fraud cases to violent crime. Several civil lawsuits against AI providers are already under way in the US, including in the context of teenage suicides.

Legally, investigators are breaking new ground: chat logs with an AI system have rarely been tested as evidence, and the question of what responsibility providers bear for minors' use of their systems remains open. The trial is therefore likely to matter well beyond the individual case.

For the AI-safety debate, the case is a wake-up call on how youth protection, reporting duties and privacy should be balanced in chatbots — a regulatory gap that even the EU AI Act addresses only partially.