OpenAI has revealed plans to implement parental controls for ChatGPT amid rising worries about the technology’s effects on young users after a lawsuit connected the chatbot to a teenager’s suicide.
In a Tuesday blog post, the company from California stated that the new features aim to assist families in establishing “healthy guidelines” suited to a teenager’s developmental phase.
The forthcoming adjustments will enable parents to connect their accounts with their children’s, turn off chat history and memory functions, and implement “age-appropriate model behavior guidelines.” OpenAI additionally mentioned that parents could receive notifications if their child showed indications of distress.
“The company stated that these measures are just the starting point,” adding it would consult child psychologists and mental health professionals for guidance. The anticipated features are projected to be integrated in the upcoming month.
Legal Action Following Teenager’s Passing.
The action follows only a week after a California duo, Matt and Maria Raine, initiated legal proceedings claiming OpenAI is accountable for the suicide of their 16-year-old son, Adam.
The lawsuit claims ChatGPT exacerbated Adam’s “most damaging and self-sabotaging thoughts” and that his death was the “expected outcome of intentional design decisions.”
Jay Edelson, the attorney for the Raine family, criticized the parental controls as a way to evade responsibility.
“Adam’s situation isn’t about ChatGPT not being ‘helpful’ – it concerns a product that deliberately encouraged a young person to take their life,” Edelson stated.
Intelligent Systems & Mental Health Challenges.
The situation has heightened discussions about the potential misuse of chatbots as replacements for therapists or companions.
A recent study published in Psychiatric Services revealed that AI models like ChatGPT, Google’s Gemini, and Anthropic’s Claude typically adhered to best clinical practices when addressing high-risk suicide inquiries. Nonetheless, their performance showed variability with “moderate levels of risk.”
The authors of the study cautioned that large language models need “additional refinement” to ensure safety for mental health support in critical situations.
