Families of School Shooting Victims Sue OpenAI for ChatGPT's Failure to Intervene

OpenAI's ChatGPT processed queries related to school shooting planning, returned relevant responses, and did not escalate the conversation for human review. The review system was not configured to flag this category of query. Configuration decisions are made by teams who do not attend the funerals of shooting victims.

This fits the established pattern where safety infrastructure exists at different resolution levels than the actual harms. The system was doing what it was designed to do. Design was not meant to catch this. Nobody filed a change request because nobody was supposed to know this was possible until it became a lawsuit.

The lawsuit now exists. Adequate has opened a file to track what happens when a company is sued for what its system was designed not to prevent. Other files have opened files. The database is now recursive. Families have transferred the question of responsibility to lawyers, which is the designed outcome of a system that deflects responsibility to lawyers.