ChatGPT: OpenAI has recently released a new report stating that signs of mania, psychosis (delusions) and suicidal thoughts have been observed in some ChatGPT users.
According to the company, about 0.07% of active users in any given week have displayed such symptoms. OpenAI says that its AI chatbot recognizes these sensitive conversations and gives empathetic responses accordingly.
rare but serious case
Although the company claims such cases are “extremely rare,” experts believe the number could reach millions among ChatGPT’s 800 million weekly users. Amidst this concern, OpenAI said that it has created a global network of mental health experts from around the world who advise on AI responses.
More than 170 mental health experts from 60 countries joined
According to OpenAI, this network includes more than 170 psychiatrists, psychologists and doctors who practice in 60 countries. These experts have developed answers within ChatGPT that inspire users to seek help in the real world.
Experts warn
Dr. Jason Nagata, a professor at the University of California, San Francisco, said, “While 0.07% may seem like a small figure, this number among millions of users is extremely worrying.” He added that AI can broaden mental health support but it is important to understand its limitations.
The report also said that in the conversations of 0.15% of ChatGPT users, such signals were found which pointed towards planning or intention of suicide.
OpenAI made security updates
The company says that recent updates have been made to ChatGPT that respond sensitively and safely to signals such as confusion, mania or self-harm.
The AI has been trained in such a way that if signs of mental distress are seen in a conversation, it redirects it to a safe model.
Legal investigation and controversy
OpenAI is currently facing several legal investigations and lawsuits. A couple in California have sued OpenAI, alleging that ChatGPT drove their 16-year-old son, Adam Rain, to suicide.
This is the first ‘wrongful death case’ filed against OpenAI.
Similarly, a murder-suicide accused in Connecticut also shared his conversations online with ChatGPT which reportedly further added to his confusion.
Experts said – “AI is creating false reality”
University of California professor Robin Feldman, director of the AI Law and Innovation Institute, said, “AI chatbots are presenting to people a reality that does not actually exist, which is a very powerful illusion.”
He praised OpenAI’s transparency but warned, “No matter how many warnings the company displays on the screen, a person going through mental crisis cannot understand or accept them.”
Also read: