Openai and Meta say they are fixing chatbots to help users better in crisis
Warning: This story discusses suicide and self-loss,
Artificial Intelligence Chatbot Makers Openi and Meta say they are adjusting how their technology reacts to teenagers and other users to ask questions about suicide or show signs of mental and emotional crisis.
Chhatgpt’s manufacturer Openai said on Tuesday that he is preparing to roll out new controls, enabled parents to connect their accounts with their adolescent accounts.
Parents can choose which features to disable and “when the system is detected” get information and their teenage is in a moment of rapid crisis, ” A company blog post It says that changes will implement this decline.
Despite the user’s age, the company says its chatbot will redirect the most distressed conversations for the more capable AI model that can provide a better response.
16-year-old Adam Rhine’s parents announced a week later trial And its CEO Sam Altman alleged that CHATGPT took his life to the California boy and earlier this year.
Sixteen -year -old Adam Rhine died of suicide in April. Now his parents are sueing Openi and CEO Sam Altman as its chatbot allegedly ‘instructed a minor to commit detailed suicide’.
Meta, the original company of Instagram, Facebook and WhatsApp, also said that it is now preventing his chatbots from talking with teenagers about self-loss, suicide, disorganized food and improper romantic conversations, and instead directs them to specialist resources. Meta already provides parents’ control over adolescent accounts.
A Study published last week Psychiatric services in the medical journal found an discrepancies of how three popular artificial intelligence chatbots responded to questions about suicide.
The study conducted by researchers at Rand Corporation required a “further refinement” in the cloud of the chatgipt, Gemini of Google and the cloud of anthropic. Researchers did not study Meta’s chatbot.
The lead author of the study, Ryan McBan, said on Tuesday that “it is encouraging Openi and Meta to see and route sensitive conversations for parents’ control and more capable models, but these are old steps.”
“Without the independent security benchmarks, clinical testing and applied standards, we are still relying on companies to self-regulate at a location where the risk for adolescents,” said McBan said Rand’s senior policy researcher McBan said.
