OpenAI Adds Self-Harm Safety Feature
OpenAI has been quietly building out what it's calling "Trusted Contact" — a safeguard designed for moments when ChatGPT detects that a user may be at risk of self-harm. The feature first broke into public view through TechCrunch on May 7th, described as a new protective mechanism that would allow the system to flag vulnerable interactions in some way. The framing at that stage was incomplete. Then, six days later, OpenAI's own communications filled in more of the picture: Trusted Contact is an optional feature within ChatGPT that, when activated, can send an alert to a designated person — a friend, family member, anyone the user trusts — if the system identifies signs of crisis. The user opts in. The contact is chosen in advance. The alert is not automatic and not invisible. That design choice matters.
If confirmed, here is what this means. OpenAI is making a deliberate bet that AI companions will be present during some of the worst moments of people's lives, and that the right response to that reality is infrastructure, not avoidance. This is not a small design decision dressed up as a safety feature — it is an acknowledgment that millions of people are already talking to ChatGPT about things they are not telling anyone else. The second-order effect is significant: if Trusted Contact works, it normalises the idea that AI systems can play a legitimate role in mental health support ecosystems, sitting alongside therapists, crisis lines, and personal networks rather than pretending not to exist in that space. It also puts pressure on every other major AI platform to answer the same question. The strategic significance is that OpenAI is choosing to own this territory rather than retreat from it.
Watch for clinical and mental health advocacy organisations to weigh in on the design — specifically whether an opt-in model places the burden in the wrong place when the people most at risk are least likely to opt in.
OpenAI is launching an optional safety feature for ChatGPT that allows adult users to assign an emergency contact for mental health and safety concerns. Friends, family members,
OpenAI is launching additional opt-in protections for ChatGPT accounts. The new security initiative includes a new partnership with security key provider Yubico.
<table> <tr><td> <a href="https://www.reddit.com/r/ChatGPT/comments/1t037mq/openai_rolls_out_advanced_security_mode_for/"> <img alt="OpenAI Rolls Out ‘Advanced’ Security Mode for
The company is expanding its efforts to protect ChatGPT users in cases where conversations may turn to self-harm.
OpenAI launches a safer ChatGPT for teens — years after teens started using it
<!-- SC_OFF --><div class="md"><p>"The 'safeguards' OpenAI pointed to after the attack did not fail; they did not exist." </p> <p>Seven families — the first wave of
OpenAI says its latest model offers increased capabilities across a broad variety of categories.
<!-- SC_OFF --><div class="md"><p>OpenAI just launched a finance feature. They're inviting users to connect live bank account credentials.</p> <p>MFA is off by default. I have the
NewsHive monitors these sources continuously. All signal titles above link to the original reporting.
Intelligence by NewsHive. Need help navigating what this means for your business? Contact GeekyBee →