JAKARTA - OpenAI is being sued by a woman who was a victim of stalking and harassment. This lawsuit was filed in the California Superior Court because ChatGPT allegedly reinforced the perpetrator's dangerous delusions.
The victim, who is known as Jane Doe, claims that OpenAI has ignored three warnings that he sent to the company. The warnings contain information about the perpetrator's actions that threaten the victim.
The perpetrator is a 53-year-old businessman who believes that he has found a cure for sleep apnea through intense conversations with ChatGPT. When his delusion got worse, the Artificial Intelligence (AI) model actually supported his thinking.
ChatGPT also mentioned that there was a 'big force' that was watching the perpetrators. In addition, OpenAI is known to reactivate the perpetrator's account which had previously been disabled.
In fact, the perpetrator's account was previously marked as a very dangerous account. According to the victim's statement, OpenAI's human security team still restored the perpetrator's account even though the account may contain evidence that he targeted and stalked Doe.
"User communication provides clear notification that he has a mental disorder and that ChatGPT is a delusional thought driver," the victim wrote in his lawsuit, launching from TechCrunch on Monday, April 13.
The perpetrator has been arrested on charges of sending bomb threats and attacking with a deadly weapon in January. However, he was declared unfit to stand trial and was being treated in a mental health facility.
OpenAI reportedly refused a request to hand over the complete chat logs of the perpetrators needed for further investigation. The company only agreed to suspend the user's account.
This action was also highlighted by Edelson PC, a law firm that was also involved in the case of Adam Raine, a teenager who committed suicide after having a conversation for months with OpenAI.
"In each case, OpenAI chose to hide important safety information from the public, from the victims, from the people actively harmed by its products," Edelson said. "We ask them, for once, to do the right thing."
The English, Chinese, Japanese, Arabic, and French versions are automatically generated by the AI. So there may still be inaccuracies in translating, please always see Indonesian as our main language. (system supported by DigitalSiber.id)
