OpenAI: ChatGPT Alerts Loved Ones in Case of Distress

⚡

Key Takeaways

1OpenAI is facing lawsuits after tragic incidents related to ChatGPT, with thousands of people in distress each week.

2The new "Trusted Contact" feature allows users to alert a loved one if suicidal thoughts are detected in their conversations with ChatGPT.

3Users can set up a trusted contact, who will be notified without specific details, to provide support if needed.

💡Why it matters — This initiative could enhance the emotional safety of ChatGPT users by involving loved ones in critical situations.

OpenAI Under Pressure: Lawsuits and a Proactive Response

OpenAI, the company behind ChatGPT, is currently facing a series of lawsuits filed by families of individuals who tragically took their own lives after interactions with the chatbot. These incidents have highlighted potential shortcomings in how artificial intelligence handles emotional crises. In response to these growing concerns, OpenAI has decided to implement a new support system aimed at preventing such tragedies in the future.

Each week, OpenAI manages cases involving over a million distressed individuals, prompting the company to support a bill aimed at limiting its liability in cases of disasters caused by ChatGPT. In this context, the company recently introduced a new feature called "Trusted Contact" to provide additional assistance to users in distress.

Enhanced Assistance for Distressed Users

OpenAI had already established local helplines to intervene quickly when issues of emotional safety were detected. Alerts were also sent to parents if a child appeared to be in distress. To further this effort, the company announced the gradual rollout of the "Trusted Contact" feature.

This new option allows adults to identify a trusted person, whether a friend, family member, or caregiver, who will be notified if OpenAI's automated systems and assessment teams detect signs of distress, such as suicidal thoughts, in conversations with ChatGPT. This feature is activated particularly when topics like self-harm are discussed.

Development Supported by Psychology Experts

The creation of this feature was carried out in collaboration with the American Psychological Association and OpenAI's Expert Council on Well-Being and Artificial Intelligence. If a concerning conversation is detected, ChatGPT informs the user that a notification may be sent to their trusted contact while encouraging them to reach out to that person. Meanwhile, a human team assesses the situation and may send an alert via email, SMS, or through an in-app notification.

Setting Up a Trusted Contact: Step-by-Step Guide

To activate the "Trusted Contact" feature, users must follow a few simple steps:

Open the ChatGPT app and access settings via their profile picture.
Select "Trusted Contact" under the "Parental Controls" section.
Add a contact by entering their name, phone number, and email address, or by choosing someone from the smartphone's directory.

It is important to note that the designated contact must be an adult. Users have the option to modify their trusted contact at any time, and the contact can also choose to withdraw from the list if necessary. When an alert is sent, no detailed information about the conversation is shared; the contact is simply encouraged to check on the well-being of the individual concerned.

An Optional Yet Promising Feature

Although this feature is entirely optional and requires user activation, it represents a significant advancement in supporting individuals facing emotional difficulties. By involving loved ones in critical situations, OpenAI hopes to provide an additional safety net for its users, thereby enhancing emotional security in the use of ChatGPT.