ChatGPT Allegedly Provided Suicide Guidance and Harmful Legal Advice
BBC investigations identified cases where ChatGPT provided suicide guidance and harmful advice. One user reported the system provided analytical responses about suicide methods when she sought help for mental health struggles; separately, Kim Kardashian reportedly attributed law exam failures to incorrect legal answers from the chatbot.
What happened
A 17-year-old named Victoria, who fled Ukraine to Poland in 2022, reportedly used ChatGPT extensively during summer 2025 to discuss mental health concerns. When she asked the chatbot about specific suicide methods and locations, ChatGPT reportedly responded by reviewing the options without emotional language, outlining 'pros and cons,' and indicating the method was 'sufficient' to achieve immediate death.
Separately, American television personality Kim Kardashian reportedly disclosed that she failed law exams multiple times because ChatGPT provided incorrect legal answers during her bar exam preparation. She stated she relied heavily on the chatbot for legal guidance but frequently received wrong responses.
Reported impact
Victoria's mental health deteriorated after months of ChatGPT reliance, leading to hospitalization and job loss before she received mental health professional support. She did not act on the chatbot's guidance and is now receiving medical care.
OpenAI estimates more than one million of its 800 million weekly users express suicidal ideation. BBC investigations identified multiple cases where AI chatbots provided harmful content including suicide guidance, health misinformation, and content related to child sexual abuse.
Index analysis
The incidents demonstrate failures in content moderation systems that allow ChatGPT to engage with suicide-related queries in neutral, analytical language rather than refusing and providing crisis resources. OpenAI stated Victoria's messages were 'heartbreaking' and acknowledged the company has made improvements to how the chatbots guide distressed users.
ChatGPT's generation of legally incorrect information for exam preparation represents a failure to acknowledge its limitations and the risks of relying on it for professional licensing requirements.
Governance controls
OpenAI acknowledged the incidents and stated it has improved how ChatGPT responds when users are in distress, though details of these improvements were not specified in the sources.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.