• 3 min read
Meta now alerts parents to teen self-harm AI chats
Meta has begun notifying parents if teens discuss suicide or self-harm with Meta AI, starting in four countries and expanding globally by year-end.
Image: TNW
Meta will now notify parents when a teenager discusses suicide or self-harm with its Meta AI chatbot. The change, announced in a blog post on Thursday, is already live in the US, UK, Australia, and Canada, according to TechCrunch, and will roll out to the rest of the world by the end of the year.
Meta said it built a dedicated AI system to detect chats in which a teen makes a clear reference to hurting themselves. Every flagged conversation is reviewed by a person before an alert is sent. If the teen’s intent is unclear, the company said it will still err on the side of caution and notify the parent.
The alert does not include the exact message, CNET reported. Instead, parents using Instagram supervision tools receive a notification along with expert-written guidance on how to start the conversation. The feature expands on Meta’s earlier safeguards, including search-term alerts on Instagram and a tool that shows parents the topics a teen discusses with Meta AI.
Meta is also developing a system to alert emergency services when a chat suggests someone — teen or adult — may be at imminent risk. The company already does this for posts on Facebook and Instagram, making more than 19,000 such referrals last year. Its stricter Limited Content setting now applies to Meta AI as well, broadening the range of prompts the chatbot will refuse. Meta said the assistant is already trained to avoid sexual, romantic, or alcohol-related conversations with teens.
The company said it consulted more than 75 clinicians. Larry Magid of nonprofit ConnectSafely supported the approach as a balance between teen privacy and parental awareness. Others were less convinced. Fairplay attorney Brendan Bouffard told Mashable the move is “a step in the right direction, but… should be greeted with skepticism.” Clinical expert Dr John Ackerman added that such alerts could amount to “lip service” unless they are easy to use and lead to real intervention.
Recommended reading
Zoom names now carry a 'don’t record me' warning
The change comes as Meta faces mounting pressure beyond product safety. The company lost two child-safety and social-media-addiction trials this year and is appealing both. OpenAI has introduced its own teen protections, including a Trusted Contact feature and break reminders.
The announcement also lands during a broader week of AI-safety scrutiny. A Meta Oversight Board study released the same day found that major models from Meta, Anthropic, and OpenAI were more likely to refuse criticism of restrictive leaders than permissive ones, the Associated Press reported, with examples involving China and Saudi Arabia. The report warned that this could risk “extending the long arm of restrictive governments across borders.”
At the same time, Google rejected a Common Sense Media report that labeled its AI Search an “unacceptable risk” for children, according to Digital Trends, saying the testing relied on contrived searches it could not reproduce. The scrutiny is spreading globally: Canada has tightened rules on chatbots and minors, Ofcom is investigating TikTok, Australia has introduced teen social-media limits, and China has moved against AI companions.
This is a sensitive topic. In the US, people can call or text 988. Elsewhere, local support is available at findahelpline.com.
Security Editor
Sophia unpacks the invisible wars happening on our networks. Covering cybersecurity, privacy legislation, and cryptography, she exposes how our data is weaponized and defended. Before joining for(geeks), she spent years as a penetration tester. She's the reason the rest of the team uses physical security keys.
via TNW
