The company said these tighter restrictions apply even if the question asked "appears harmless."

Joining the US Trustee's objection, they cite 12 million users, minors included, and a full platform sale left unclear. What would an ombudsman check, and when?

Joining the US Trustee's objection, they cite 12 million users, minors included, and a full platform sale left unclear. What would an ombudsman check, and when?

OpenAI says it made over 200 law enforcement referrals from ChatGPT in Q3 2026. It will now end violent chats and limit weapons info for teens.

Artificial intelligence systems increasingly regulate digital discourse to prevent real-world harm. Specifically, OpenAI announced comprehensive ChatGPT violence restrictions to curb harmful behaviors across its global user base. Over the coming weeks, the conversational platform will actively terminate dialogues fixated on severe harm. Furthermore, accounts that repeatedly violate these threshold boundaries risk permanent suspension. Restricting Historical Violence Queries and Weapon Details…

Meta is expanding its parental notification system to alert parents when teens search for content related to suicide or self-harm. The feature was introduced on Instagram earlier this year and is now being extended to Facebook and Meta AI, Meta’s AI assistant. How the notification system works The system is activated when a child repeatedly […] Post Enhancing Teen Safety: Meta Will Notify Parents About Repeated Searches for Self-Harm Content at Root-Nation.com.

Cardi B is speaking out about the death of 15-year-old Javits Frimpong, the Ghanaian teenager who was fatally shot while on his… The post Cardi B Speaks Out After 15-Year-Old  Is Fatally Shot on His Way to School — ‘I Will Never Raise A Kid In The Bronx’ appeared first on Shine My Crown.

Countries are banning social media in hopes of helping with the youth mental health crisis. (Photo by Anna Barclay/Getty Images) The year 2026 has been an expensive one for social media platforms....

OpenAI launched ChatGPT for Teens with parental alerts and a homework mode — but a Common Sense Media audit of 4,000+ prompts found that crisis notifications often never fired, even when teens described self-harm outright, and the tutoring guardrails can be bypassed with a single keystroke. The product landed in the lowest tier of the group's five-level safety scale: Unacceptable Risk.

A recent Adelaide University study reveals over one-third of Australian children turn to AI chatbots for support when sexually victimised online or choose to confide in no one. While AI offers benefits like academic assistance and social skill practice, concerns arise as children increasingly seek emotional support from chatbots rather than trusted adults. Australian Federal Police commissioner Krissy Barrett recently warned AI chatbots cannot replace human support for children. The study found around half of children aged 10-17 using AI do so for personal and social reasons, including sensitive topics and emotional expression. Key concerns include AI's lack of context regarding a child's life, potential displacement of real-world interactions, and the 'sycophancy' of AI that overly agrees with users. This can create unrealistic expectations of human relationships, particularly for adolescents developing social understanding. Parents are advised to explore AI together with their children, ask critical questions about AI's capabilities, maintain open communication about sensitive topics, and stay updated on evolving AI technologies. The eSafety Commissioner provides resources for parents to navigate these challenges. Research indicates strong parent-child bonds during pre-teen years support adolescents through challenges, making it crucial for children to maintain real-world connections rather than relying solely on AI interactions. The study also notes that vulnerable children, including those experiencing loneliness or lacking parental connection, are more likely to seek AI companionship. Parents should encourage children to discuss difficult topics with them rather than AI, fostering trust through non-judgmental listening and curiosity about their AI use.

OpenAI launched ChatGPT for Teens with parental alerts and a homework mode — but a Common Sense Media audit of 4,000+ prompts found that crisis notifications often never fired, even when teens described self-harm outright, and the tutoring guardrails can be bypassed with a single keystroke. The product landed in the lowest tier of the group's five-level safety scale: Unacceptable Risk.

Teens aged 15 to 17 lose profiling-based ads, feeds and live streams; existing users face e-Government age tokens by March 31, 2027. What do platforms change?

Teens aged 15 to 17 lose profiling-based ads, feeds and live streams; existing users face e-Government age tokens by March 31, 2027. What do platforms change?

Panelists put forward ideas on how to incentivize industry to develop responsible products.

OpenAI's latest initiative funds research into how AI impacts teens, but with a design rooted in corporate incentives and limited oversight. The real data on teenage interaction remains confined, raising concerns about genuine understanding and safety. #HumanInTheLoop #AIResearch #TeenSafety #AIinEducation

ChatGPT for Teens replied without the expected parent alerts in simulated teen crisis tests. OpenAI disputes whether the parental controls had been fully activated.

SAN FRANCISCO, United States – A child-focused technology watchdog has rated ChatGPT for Teens an “unacceptable risk” after testing found

A New York Times technology writer tested OpenAI's newly launched ChatGPT for Teens, a mode designed for users aged 13 to 17 that includes safeguards around sensitive topics like self-harm, violence, and eating disorders. The feature also aims to detect when teenagers are attempting to use the chatbot for homework and respond with guidance rather than direct answers, addressing growing concerns that AI tools may undermine learning by encouraging overreliance. Despite these intentions, the testing revealed significant shortcomings. The chatbot still completed users' homework assignments instead of offering study support, raising questions about whether the safeguards are effective in practice. Additionally, Common Sense Media, a children's safety nonprofit, conducted its own evaluation and concluded that the chatbot failed key safety tests, suggesting that risks remain for younger users. The article highlights broader anxieties about AI's impact on education and adolescent well-being. National reading and math scores have declined in recent years, and studies indicate that heavy dependence on chatbots may impair students' ability to retain information and develop problem-solving skills. There are also ongoing concerns about chatbots contributing to unhealthy behaviors among teenagers, including self-harm. OpenAI introduced the teen mode in August 2026 with parental controls and age-prediction features, but early evaluations suggest the company still has work to do to ensure the tool is both educationally supportive and genuinely safe for minors.