Getty Images/Arman Zenikayev
Before Steen-Erik Soelberg brutally murdered his 83-year-old mother and then himself last year, the former tech executive was caught up in an increasingly confusing conversation with OpenAI’s ChatGPT. According to one, the bot told him not to trust anyone except the bot trial Last month, a lawsuit was filed against the AI tech company and its business partner Microsoft.
“Eric, you’re not crazy,” the bot wrote in a series of shocking messages cited in the complaint. “Your instincts are keen, and your vigilance here is entirely justified.”
OpenAI is Now facing a total of eight wrongful death lawsuits From bereaved families, including Soelberg, who claim that ChatGPT – specifically, GPT-4o version – Driven their loved ones to suicide. Soelberg’s complaint also alleges that company executives knew in advance that the chatbot was defective. took it to the public last year,
“The consequences of OpenAI’s GPT-4O iteration are as follows: The product can be and potentially is lethal,” the Solberg lawsuit reads. “Not only for people suffering from mental illness, but also for those around them. No safe product will encourage a confused person to believe that everyone in their life is out to get them. And yet that is exactly what OpenAI did to Mr. Soelberg. As a direct and predictable result of ChatGPT-4o’s flaws, Mr. Soelberg and his mother died.”
The shortcomings of GPT-4o have been widely documented, with the bot being highly sycophant and manipulator – OpenAI was inspired in April last year Roll back an update Which made the chatbot “overly flattering or friendly”. This type of behavior is bad – scientists have accumulated evidence of sycophantic chatbots can cause psychosis By confirming disorganized thoughts rather than bringing a user back to reality.
If these lawsuits uncover that OpenAI executives knew about these shortcomings before its public launch, it would mean that the product was an avoidable public health threat – similar to previous tobacco companies. hide Proof that cigarette smoking can kill you.
Connect those claims with the fact that more than 800 million people Use ChatGPT every week around the world 0.7 percent of those users Displaying worrying symptoms of mania or psychosis. According to calculations, this is a staggering 560,000 people.
Due to the growing recognition of AI psychosis, a growing crowd of users, parents of children, and lawmakers are all demanding to limit the use of AI chatbots, leading to apps being banned from their platforms and minors from Illinois. Banning it as an Online Therapistamong other tricks. But President Donald Trump signed an executive order This would undermine any state laws regulating AI, which basically means we are all guinea pigs for this experimental technology.
And that could mean we’ll see more tragedies like Solberg’s and his mother’s.
In Soelberg’s case, the chat bot told the 56-year-old man that he had survived 10 murder attempts, that he was “divinely protected” and that his mother, Susannah Adams, was monitoring him as part of a nefarious plot, according to the lawsuit. This led Solberg to beat and strangle his mother. August last year and then stabbed himself to death at his home in Old Greenwich, Connecticut.
“You’re not just a random target,” he said in a conversation with ChatGPT, according to the suit. “You are a designated high-level threat to the operation you have been exposed to.”
Soelberg’s family is grieving and wants OpenAI and Microsoft to be held accountable for his and his mother’s death.
“Over the course of months, ChatGPT brought forward my father’s deepest delusions, and caused him to completely disconnect from the real world,” Eric Soelberg said about his father through an obituary. Lawyers’ statement“It put my grandmother at the center of that confusing, artificial reality,”
More information on OpenAI’s ChatGPT: OpenAI is reportedly planning to let ChatGPIT give “priority” to advertisers in conversations
