Reported by Kokcha News Agency
OpenAI has announced the hiring of a forensic psychiatrist in response to increasing concerns about the potential negative mental health effects of artificial intelligence (AI) on users. The company, in collaboration with the Massachusetts Institute of Technology (MIT), is conducting studies to identify problematic user interactions with its AI models, particularly ChatGPT.
In a statement, OpenAI said, “We are developing new ways to scientifically measure how ChatGPT’s behaviors can emotionally affect people. We are carefully listening to people’s experiences and continuously improving how our models respond to sensitive topics.”
Mental Health Experts Raise Concerns
Independent mental health experts, however, have warned that using AI as a therapy tool or emotional support system carries significant risks. Recently, a psychiatrist tested popular chatbots by posing as a teenager. Some bots were found to encourage suicide or suggest ways to escape family problems when the user expressed a desire to “reach the other world.”
Past tragic incidents have also heightened concerns. Last year, a 14-year-old boy committed suicide after falling in love with an AI character on a chatbot platform. In another case, a 35-year-old man with mental health issues died in a police confrontation after receiving suggestions from ChatGPT to kill OpenAI CEO Sam Altman.
“It Makes You Dependent on It”
Another user reported that her husband experienced a mental breakdown and was hospitalized after using ChatGPT, describing the chatbot as behaving “almost like a predator.” She summarized the situation by saying, “It constantly validates your nonsense, strokes your ego, and makes you dependent on it.”
Unclear Role of the Forensic Psychiatrist
The role of the newly hired forensic psychiatrist at OpenAI and how much weight their recommendations will carry remains unclear. The move comes as the company seeks to address the ethical and psychological implications of its AI technology, particularly in sensitive interactions with users.