\n<\/p>\n
OpenAI began testing a new\u00a0safety routing\u00a0system<\/a>\u00a0in ChatGPT\u00a0over the weekend, and\u00a0on Monday\u00a0introduced\u00a0parental controls\u00a0to\u00a0the\u00a0chatbot\u00a0\u2013\u00a0drawing mixed reactions from users.<\/p>\n The safety features come in response\u00a0to\u00a0numerous\u00a0incidents\u00a0of\u00a0certain\u00a0ChatGPT\u00a0models\u00a0validating\u00a0users\u2019 delusional thinking\u00a0instead of\u00a0redirecting\u00a0harmful conversations.\u00a0OpenAI is facing a\u00a0wrongful death lawsuit<\/a>\u00a0tied\u00a0to one such incident,\u00a0after\u00a0a teenage boy died by suicide after months of interactions with ChatGPT.\u00a0<\/p>\n The routing system is designed to\u00a0detect emotionally sensitive conversations and\u00a0automatically switch mid-chat to GPT-5-thinking, which the company sees as the best equipped model for high-stakes safety work. In particular, the GPT-5 models were trained with a new safety feature that OpenAI calls \u201csafe completions,\u201d<\/a> which allows them to answer sensitive questions in a safe way, rather than simply refusing to engage.<\/p>\n It\u2019s a contrast from the company\u2019s previous chat models, which are designed to be agreeable and answer questions quickly. GPT-4o has come under particular scrutiny because of its overly sycophantic, agreeable nature, which has both fueled incidents of\u00a0AI-induced\u00a0delusions and drawn a large base of devoted users.\u00a0When OpenAI rolled out\u00a0GPT-5 as the default\u00a0in August<\/a>, many users\u00a0pushed back\u00a0and demanded access to GPT-4o.\u00a0<\/p>\n While many experts and users have welcomed the safety features, others have\u00a0criticized\u00a0what they see as an overly cautious implementation, with some users\u00a0accusing OpenAI of treating adults like children in a way that degrades the quality of the service.\u00a0OpenAI has suggested that getting it right will take\u00a0time and\u00a0has given itself\u00a0a\u00a0120-day period of iteration and improvement.\u00a0\u00a0<\/p>\n Nick Turley, VP and head of the ChatGPT app, acknowledged some of the \u201cstrong reactions to 4o responses\u201d due to the implementation of the router\u00a0with explanations.\u00a0\u00a0<\/p>\n \u201cRouting happens on a per-message basis; switching from the default model happens on a temporary basis,\u201d\u00a0Turley posted on X<\/a>. \u201cChatGPT will tell you which model is active when asked. This is part of a broader effort to strengthen safeguards and learn from real-world use before a wider rollout.\u201d\u00a0<\/p>\n Techcrunch event<\/p>\n \n\t\t\t\t\t\t\t\t\tSan Francisco<\/span> The implementation of parental controls in ChatGPT received similar levels of praise and scorn<\/a>, with some commending giving parents a way to keep tabs on their childrens\u2019 AI use, and others fearful that it opens the door to OpenAI treating adults like children. <\/p>\n The controls let parents customize their teen\u2019s experience by setting quiet hours, turning off voice mode and memory, removing image generation, and opting out of model training. Teen accounts will also get\u00a0additional\u00a0content protections<\/a>\u00a0\u2013 like reduced graphic content and extreme beauty ideals\u00a0\u2013 and a detection system that recognizes potential signs that a teen might be thinking about self-harm.\u00a0\u00a0<\/p>\n \u201cIf our systems detect potential harm, a small team of specially trained people reviews the situation,\u201d per OpenAI\u2019s\u00a0blog<\/a>. \u201cIf there are signs of acute distress, we will contact parents by email, text\u00a0message\u00a0and push alert on\u00a0their\u00a0phone, unless they have opted out.\u201d\u00a0<\/p>\n OpenAI acknowledged that the system\u00a0won\u2019t\u00a0be perfect and may sometimes raise alarms when there\u00a0isn\u2019t\u00a0real danger, \u201cbut we think it\u2019s better to act and alert a parent so they can step in than to stay silent.\u201d The AI firm said it is also working on ways to reach law enforcement or emergency services if\u00a0it\u00a0detects\u00a0an imminent threat to life and\u00a0cannot\u00a0reach a parent.\u00a0<\/p>\n<\/div>\n
\n\t\t\t\t\t\t\t\t\t\t\t\t\t|<\/span>
\n\t\t\t\t\t\t\t\t\t\t\t\t\tOctober 27-29, 2025<\/span>\n\t\t\t\t\t\t\t<\/p>\n<\/p><\/div>\n<\/p><\/div>\n<\/div>\n