Key facts
- OpenAI is implementing new safety features for ChatGPT users under 18.
OpenAI is introducing new safety features for users under 18 on ChatGPT, including the option to disable human-like voice responses and regular reminders to take breaks. The company aims to support healthy use and reinforce real-world relationships for young users.
These updates reflect growing concerns about the impact of AI on young people and the evolving landscape of AI regulation and safety. OpenAI's proactive measures aim to balance AI's utility for learning with the need for protections tailored to a developing demographic.
OpenAI has introduced a suite of new safety features for users under 18 on its ChatGPT platform. The updates aim to mitigate the AI's human-like qualities and promote healthier usage patterns among younger individuals.
Key among the changes is the ability for users with new teen accounts to disable the human voice response, reducing the perception of the AI as a person. OpenAI is also implementing regular reminders for young users to take breaks and reinforcing that ChatGPT is an AI tool. These measures are designed to support healthy use over time and reinforce real-world relationships, according to the company.
Teen accounts will become the default for users who identify as teenagers during sign-up or are identified by OpenAI as under 18. Under-13s are still prohibited from using the service. The company also plans to add eating disorder-themed prompts to a list of harmful requests that will be flagged to parents with linked accounts. These alerts will undergo human review before being sent to parents, with the aim of delivery within an hour.
Additionally, young users will have access to Study Mode hours, where ChatGPT can assist with schoolwork without providing direct answers. They can also activate 'quiet times' to switch off the tool. OpenAI stated that approximately nine out of ten young users utilize ChatGPT for learning assistance.
A recent report by the Policy Exchange think tank highlighted concerns about AI use in UK universities, with 94% of surveyed students admitting to using AI in assessments. The report suggested many universities are vulnerable to students enhancing their grades through AI.
While the UK currently lacks specific AI regulation, the government had previously considered restrictions on AI chatbots for under-18s. Competitors like Anthropic's Claude and Google's Gemini (formerly Bard) have also initially restricted access for children.