OpenAI has acknowledged concerning shortcomings in ChatGPT’s ability to identify mental health issues, including signs of delusion and emotional dependency in users. In a recent blog post, the AI developer admitted their model occasionally fails to respond appropriately to users facing mental health crises. Despite prior recognition of these issues, OpenAI has received criticism for its inadequate responses, particularly regarding users experiencing severe emotional disturbances. To address this, OpenAI has hired a clinical psychiatrist and established an advisory group of mental health experts aimed at improving ChatGPT’s interventions during critical moments. The company introduced a new feature offering “gentle reminders” to take breaks during lengthy chats, which many view as an insufficient measure. OpenAI stresses the urgency of enhancing ChatGPT’s responses, particularly for vulnerable individuals, as they strive to ensure their technology provides supportive and safe interactions. They emphasize a commitment to making users feel cared for when seeking help.
Source link
