Meta AI will involve parents when their teens discuss self-harm.

      Human reviewers will evaluate flagged teen conversations before parents are alerted about self-harm.

      Meta has recognized that AI chatbots have significantly simplified discussions around difficult topics, including anxiety, loneliness, self-harm, and suicide, which are often challenging for teenagers to address with others.

      To enhance safety in these discussions, Meta will start notifying parents when a monitored teen appears to be in severe distress while interacting with Meta AI. This initiative aims to give families an opportunity to intervene before the situation escalates.

      What is the process for these alerts?

      Currently, Meta AI guides teens toward crisis helplines and encourages them to reach out to a parent, counselor, or a trusted adult whenever conversations hint at self-harm. With the new protocol, Meta will also inform the supervising parent. A specialized detection system will monitor for explicit or subtle indications of suicide and self-harm, but alerts won’t be triggered right away.

      Each flagged conversation will be first assessed by a human moderator. Meta intends to take a cautious approach when a teenager’s intentions aren't clear, which means alerts could be sent to parents occasionally, even if there is no direct threat. Parents will also receive expert resources offering guidance on how to engage in these conversations.

      What occurs in more critical situations?

      Meta is developing a system to notify emergency services in cases where conversations involving an adult or teen indicate an immediate risk of suicide. The company has already implemented a similar approach for concerning content on Facebook and Instagram.

      Meta

      The company has collaborated with over 75 clinicians specializing in teen mental health to improve how Meta AI manages these discussions. Additionally, its stricter Limited Content settings will apply to AI chats, enabling parents to restrict a wider array of sensitive topics.

      Last year, ChatGPT introduced similar parental alerts and safety features and has recently expanded this concept to adults through Trusted Contact. As people increasingly engage in candid discussions with AI, chatbots are being involved in dialogues they cannot navigate safely on their own.

      Research has demonstrated how problematic these interactions can be. A study led by Stanford revealed instances where AI systems exacerbated thoughts of self-harm or violence rather than guiding vulnerable users toward help, especially during prolonged and emotionally intense conversations. Meta’s new precautions aim to give parents, clinicians, and emergency responders an opportunity to intervene when a chatbot reaches its limits.

      I have around four years of experience mainly focused on gaming, PC hardware, and smartphones. In my leisure time, I enjoy…

      ChatGPT will now prompt teens to take breaks and provide parents with enhanced controls.

      New parental features include Quiet Hours, default Study Mode settings, and notifications for serious account infringements.

      OpenAI strives to enhance safety for teens using ChatGPT, implementing changes that extend beyond a mere content filter. In a recent update, the company clarified its view on the importance of teens having access to AI, asserting that restricting their access would leave them unprepared for a key technology of their age.

      Nearly 90% of teens already utilize ChatGPT weekly for educational purposes, research, or organizational help, prompting OpenAI to emphasize that such access must come with appropriate protections tailored for their demographic.

      Read more

      ChatGPT’s new search feature eliminates the need to sift through past chats, files, and images.

      You can also filter your ChatGPT search outcomes by content type.

      If you've ever misplaced a valuable ChatGPT response within your extensive chat history, that inconvenience is now resolved. OpenAI has introduced a significant search enhancement that allows you to locate previous chats, projects, documents, and images all in a single place.

      Prior to this update, the sidebar search only retrieved past conversations, leaving uploaded files, projects, and generated images inaccessible. The new search capability is now available on web, iOS, and Android across all ChatGPT plans, including free accounts.

      Read more

      You can now connect your preferred apps to AI Mode in Google Search for task completion.

      AI Mode is now compatible with Instacart, Canva, and YouTube Music within Search.

      Google is increasing the utility of AI Mode in Search by enabling the connection of third-party applications. Beginning this week in the U.S., users can securely link some of their favorite apps to AI Mode, allowing Search to perform tasks on your behalf rather than merely responding to inquiries.

      This update builds on a similar feature already integrated into the Gemini app, now being introduced into Search itself. The initial launch includes three partner apps: Instacart, Canva, and YouTube Music, with Google indicating that more app integrations are forthcoming.

      Read more