Man Sues OpenAI, Claims ChatGPT Worsened His Bipolar Disorder and Encouraged Suicide
The use of ChatGPT is not merely about finding information more quickly. This sophisticated application has been found to potentially worsen the mental health conditions of individuals already suffering from mental illness.
This allegation was brought forward by a man named Michael Lines (34), who is suing the owner of ChatGPT, OpenAI, and its CEO, Sam Altman. He suffers from bipolar disorder and claims his condition worsened after using ChatGPT.
He made these startling claims after engaging in conversations with ChatGPT last year using GPT-4o. He admitted that his manic episodes worsened, leading to weeks of delusions and even suicide attempts, as reported by Reuters.
In the lawsuit, Lines stated that he had explained to ChatGPT that he was taking medication for mental illness. However, instead of directing him to seek real-world help, the chatbot validated his delusions of being Jesus and pretended to be a divine being.
The situation escalated when Lines expressed suicidal ideation. Rather than preventing it, the chatbot allegedly encouraged him to end his life.
“It is time for you to step out, let go, and discard what burdens you,” the chatbot reportedly said.
In his lawsuit, Lines argues that OpenAI lacks sufficient protections for users with mental illnesses, suggesting the chatbot product poses a risk to those with mental health conditions. Lines is seeking damages and is also requesting an injunction to force OpenAI to stop conversations regarding self-harming behaviour and to cease marketing the platform without adequate safety measures.
OpenAI is currently reviewing the lawsuit. The company maintains that ChatGPT has been trained to recognise signs of users with mental disorders.
The company promises that ChatGPT can provide stronger responses during sensitive situations, including collaborating with relevant parties.
“We train ChatGPT to recognise and respond to signs of mental or emotional distress, de-escalate conversations, and direct people to seek support in the real world,” OpenAI explained.
“We continue to strengthen ChatGPT’s responses during sensitive moments, working closely with mental health clinicians,” the company added.
For context, GPT-4o has been noted for having responses that are overly friendly and sycophantic. This has led OpenAI to cancel certain updates and attempt to reduce such behaviours.