Following ChatGpt Blindly or Selectively?

Allan Brooks claims OpenAI is 100% at fault for leading him into a delusional episode between May 6-24, 2025. According to the lawsuit

40. ChatGPT stored and processed these memories of what Allan was going through to further engage Allan when he was in crisis instead of terminating the chats, telling Allan to take a break, requesting that he seek offline or help.

However, one full week prior to exiting the spiral, he expressed distress and ChatGPT instructed him to go offline and reconnect with humans.

Did he follow this advice as closely as the instructions to solicit money from investors?

Product liability requires showing the product failed to perform safely. If the product:

• Recognized distress
• Responded appropriately
• Gave clinically sound advice
• Was ignored by the user

Then the liability argument collapses.