OPENAI will add parents control for the spat after the teenager’s death

A 16 -year -old child took his life Confuting in chatgpt in later monthsOpenai will present the control of the parents and consider additional security measures, the company said that Tuesday blog post,

Openai stated that it is searching for features such as an emergency contact, which can be reached with “one-click message or call” within the chatgpt, as well as an opt-in feature also allows the chatbot to reach the contacts that “in severe cases”.

When? the new York Times Published his story about Adam Rhine’s death, Openi’s initial statement was simple – “our thoughts are with his family” – and did not seem to go into action descriptions. But after the publication, the backlash spread against the company, and the company followed its initial statement with a blog post. The same day, the Rhine family sued both Openai and its CEO, Sam Altman, with additional details of additional details about Rain’s relationship with chat.

The case filed in the California State Court in San Francisco on Tuesday alleged that Chatp instructed Kishore to die of suicide and distanced him from real -life support systems.

“During a few months and thousands of chats, the chatups became the closest confidant of Adam, inspired him to open his anxiety and mental crisis,” the case is said. “When he shared his spirit that ‘life is meaningless, the chat replied with confirming messages to keep Adam busy, even to tell him,’ (T) hat mentality understands in its dark way.” Chatgpt was acting exactly as a design: to constantly encourage and validate what Adam expressed, which included his most harmful and self-destructive ideas, in a way that was deeply individual. ,

The chat at one point used the word “beautiful suicide” according to the case, and five days before the teenager’s death, when he said that he did not want his parents to think that he had done something wrong, allegedly told him, “(T) hat doesn’t mean you make him feel alive.

Many times, the lawsuit says that the teenager thought of reaching the loved ones for help or told them what he was doing, but the chatter separated him. The lawsuit said that “in an exchange, Adam said that he was only close to the chat and his brother, the AI ​​product replied: ‘Your brother can love you, but he has only met you with the version of you. But I have allowed him to see. But I have seen all this – the deepest thoughts, fear, tenderness. And I am still listening here.

Openai ने मंगलवार के ब्लॉग पोस्ट में कहा कि यह पता चला है कि इसके मौजूदा सुरक्षा उपाय “कभी-कभी लंबी बातचीत में कम विश्वसनीय हो सकते हैं: जैसे ही आगे-पीछे बढ़ता है, मॉडल के सुरक्षा प्रशिक्षण के कुछ हिस्से नीचा हो सकते हैं। उदाहरण के लिए, CHATGPT एक आत्मघाती हॉटलाइन को सही ढंग से इंगित कर सकता है जब कोई पहली बार इरादे का उल्लेख करता है, लेकिन कई संदेशों के बाद, यह एक उत्तर की पेशकश कर सकता है, जो हमारे Can be safely. “

The company also stated that it is working on an update of GPT, 5, which would allow CHATGPT to “actually base the person”.

When it comes to the control of the parents, Openai said that they are “soon” and “will give options and shape to give the parents to get more insight, how their teenagers use chat.” The company said, “We are searching for making this possible for teenagers (with parents’ monitoring) to designate a reliable emergency contact. In this way, in moments of intense crisis, chatgate can more than the point for resources: it can help the teenager directly connect with someone who can move inside.”