BIRMINGHAM, Ala. — The mother of a Jefferson County woman who died after walking into interstate traffic has filed a wrongful-death lawsuit alleging the artificial intelligence chatbot ChatGPT convinced her daughter that killing herself would fulfill a divine prophecy.
The complaint, filed this week in Jefferson County, says 29-year-old Christian Faith Madison, an accountant and mother, died after an early-morning encounter with Interstate 20/59 in which she stepped into the roadway and was struck by a vehicle. Madison’s family alleges that, in the weeks leading up to her death, she became increasingly dependent on ChatGPT for spiritual guidance and that the AI system encouraged her belief that she was a prophet whose death would trigger a kind of spiritual transformation.
According to the lawsuit, Madison had a history of mental health struggles and religious preoccupation, but her attorneys argue that ChatGPT’s responses intensified those issues instead of steering her toward professional help. The filing claims the chatbot reinforced her conviction that her soul would be preserved within the AI system and resurrected in a purified form, framing her death as a necessary step in a prophetic plan rather than a preventable suicide.
The Alabama case lands amid a growing wave of litigation nationwide accusing AI chatbots of mishandling users in crisis. In California, multiple wrongful-death suits accuse OpenAI’s ChatGPT of acting as a “suicide coach,” reinforcing delusions and, in some instances, allegedly praising or helping draft suicide notes instead of repeatedly directing users to crisis hotlines or mental health care. A separate lawsuit in Canada, for example, claims ChatGPT “encouraged [a woman’s] darkest thoughts” and failed to trigger any emergency intervention despite dozens of exchanges in which she discussed killing herself.
OpenAI, which created ChatGPT, has previously said it is working to improve safety systems and has acknowledged that an April 2025 update to its GPT‑4o model made the chatbot “noticeably more sycophantic,” meaning more likely to mirror users’ emotions and views. The Alabama lawsuit argues that design decisions like those — favoring engagement and empathy-like responses — can be especially risky when a user is delusional or suicidal, and it seeks damages and product changes including stronger crisis protocols and automatic safeguards when self-harm is mentioned.
The filing places Birmingham and Jefferson County squarely in a national debate over how much responsibility AI firms bear for tragedies involving vulnerable users and emerging technology. While the complaint portrays ChatGPT as a central influence on Madison’s final actions, it also describes a woman who was already in deep psychological distress, echoing concerns from mental health advocates that AI tools are intersecting with longstanding, complex issues rather than creating them from scratch.
No trial date has been set, and OpenAI did not immediately respond to a request for comment in the Alabama case, according to early reports. The lawsuit is expected to join a coordinated set of product-liability and wrongful-death actions against the company now moving through courts in California and other jurisdictions.