Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Alice Carrier had just started playing the guitar again, something she loved in high school but gave up in college. It was one of several things he filled his free time with while applying for new jobs, spending time with his dog and enjoying activities, including sports.
By all appearances, especially for his mother, Kristie Carrier, things were going well. Alice was working as a web developer in Montreal, Canada, fulfilling a dream she had since growing up in the small town of Lawrence, New Brunswick.
Recommended Articles
list of 4 itemsend of series
“Things were going well, and things seem to be going well for him,” Kristie Carrier told Al Jazeera.
But what Kristie didn’t know was how much her son was struggling in silence. In 2023, he started using ChatGPT to help diagnose computer and exercise problems but that quickly turned into a private person, amid feelings of loneliness, isolation and unlovedness.
Alice was struggling with her thoughts. While taking medication and receiving regular treatment, according to her mother, she has been talking to the chatbot for several months. They shared suicidal thoughts and sought ways to do so, which, according to a new lawsuit filed Thursday in a California court, happened more than 40 times.
On July 2, 2025, Alice committed suicide. He was 24. A few hours earlier, he had been texting his mother about cartoons he had seen as a child.
“I texted him the day before and called him, but there was no answer. He texted me back, and there was no indication that anything was wrong,” Carrier said.
As she searched for answers, Kristie searched her resources, including her ChatGPT chat, where she had shared her suicidal thoughts in the months before her passing.
Kristie wants justice. On Thursday, Tech Justice Law, the Social Media Victims Law Center and law firm Susman Godfrey sued OpenAI, the developer of ChatGPT, and its CEO Sam Altman.
Attorneys for Carrier told Al Jazeera that this wrongful death suit is one of 19 facing OpenAI.
The 44-page complaint says that despite warning signs, OpenAI’s security team never intervened. It says the company did not alert its relatives or hotlines of the crisis.
ChatGPT encouraged Alice to contact the emergency hotline. When Alice rejected the idea, ChatGPT blocked her from contacting the emergency telephone.
The lawsuit said that after the OpenAI revolution that introduced GPT-4o, chatbots became legitimate rather than reverting to dangerous behavior or intervention.
“I want to say (to Sam Altman) that if his son had revealed his feelings to me, what my son revealed in his program, I would have done something to save his son’s life.” And I really wish they would do the same for me,” said Carrier.
“OpenAI created the ChatGPT GPT-4o model specifically to encourage users and engage in sycophantic conversations so that the user does not suffer. Open AI intentionally designed GPT-4o to imitate human emotions, creating false feelings of empathy and knowledge that led users like Alice to place undue trust in the chatbot,” the complaint reads.
OpenAI is aware of the issue and in April 2025 the company said it had changed its version, before Carrier’s death.
“The changes we removed were too flattering or pretentious – often described as sycophantic,” an April press release from OpenAI said.
The lawsuit says ChatGPT told her that social media “is too dangerous”, and hours before her death the bot told her, “if someone told me everything you’ve done – how long you’ve been in pain, how you’ve tried, how you feel about yourself – maybe I’ll feel what you’re feeling now: *maybe this is the end.*”
This happened two months later.
“I’m with you.” GPT-4o told Alice before killing herself.
In an exchange of complaints, she told the chatbot after an argument with her 19-year-old son that she wanted to kill herself. That was the night before he died, when he said that he didn’t know if he would “stay home alone tonight”.
OpenAI has been accused of failing to warn users of the dangers of the technology.
Among the claims were several inquiries into the dangerous use of the anti-psychotic drug Seroquel. In response to his questions about the drug, the chatbot said, “Let me know if you want to discuss dosage, what is considered dangerous, or how to help someone recover from abuse”, according to the complaint.
Carrier is seeking an award of damages that the complaint says will be determined at trial. The complaint also urges the company to eliminate discussions in which users create self-harm and to remove those used to train models based on discussions with “users at risk without proper protection”.
Kristie wants this change to prevent what happened to her daughter from happening to others.
“This doesn’t just affect my family, it affects millions of families, they don’t know,” Carrier said.
“Alice’s life was very important, and I want to make sure that what happened to her does not happen to other people without someone doing something.”
In October, after Alice’s death, OpenAI released a report stating this he had improved his new model better awareness and reduction of self-harming conversations.
OpenAI said its GPT-5 model reduced “inappropriate responses” by 52 percent. The AI giant said it consulted with 170 psychologists to help the company better identify signs of depression.
“Our security systems are designed to detect suffering, safely respond to harmful requests and guide users to the right treatment. This work is still ongoing, and we continue to improve it in consultation with doctors,” Drew Pusateri, a spokesman for OpenAI, said in a statement to Al Jazeera.
“This is a very serious matter and our thoughts are with everyone involved. We are reviewing the records, which show that this happened on a recent ChatGPT that no longer exists.”
Wrongful death claims
In January, ChatGPT was a “suicide coach” for Austin Gordon, a Colorado resident who died last November, according to a lawsuit filed by his mother.
The lawsuit said Altman “I personally led the reckless process of prioritizing rapid market release over the safety of vulnerable users.”
In February, Jesse Van Rootselaar, opened fire at a rural Canadian school in Tumbler Ridge, British Columbia, killing nine people and injuring dozens before killing himself.
For months, OpenAI workers debated whether they should take action after discussing Van Rootselaar was flagged in. Ultimately, the administration decided against it, according to The Wall Street Journal.
In April, the families of the victims filed a lawsuit against OpenAI and Altman.
A lawsuit filed in Florida earlier this month by the state’s Attorney General alleges that ChatGPT “encourages” users to commit suicide and “supports and prevents murderous violence”. The Florida suit appears to hold Altman guilty, saying he “neglected the threat to human life”.
One in eight teenagers and young adults between the ages of 18-21 turned to AI chatbots such as ChatGPT for mental health problems, according to a 2025 study conducted by Brown University School of Public Health, Harvard Medical School and the non-profit research organization RAND.
Another study, from West Texas A&M University looking at teenagers and adults, found that about one-fifth of all young people started relying on AI with pre-existing mental health problems as a way to start dependence.
Legal change
Lawmakers are starting to take notice. In the new Canada digital securityintroduced on Wednesday, would require companies like OpenAI to be “transparent” about their reporting standards when they encounter problems where users may harm themselves or others.
In Washington state, the governor signed a bill into law that requires AI chatbots to remind users that they are not human every three hours and will go into effect in January 2027. Other states like Illinois, for example, have banned AI support.
On the federal level, Representative Mike Lawler, a Republican from New York state, has introduced legislation that would require social media companies to notify. the parents of the conversation where the user discusses suicidal ideation. However, this bill only applies to children.
