OpenAI released its first report on teen ChatGPT use on 7 October, saying teenagers spend under 15 minutes a day on the service on average and fewer than 2% use it for more than three consecutive hours, Reuters reported. The figures arrive on the same day that Common Sense Media called OpenAI’s teen safeguards an “unacceptable risk” following its own tests, Futurism reported.
Key points
- OpenAI says average daily teen use is under 15 minutes.
- The company says fewer than 2% of teens use ChatGPT for more than three consecutive hours.
- Common Sense Media says its testing found gaps in the product’s safety features.
OpenAI puts a number on teen use
OpenAI introduced ChatGPT for Teens on 18 August as the default experience for accounts identified as belonging to users under 18, Unite.AI reported. The tier includes Study Mode, quizzes, Learning Visualizations and scheduled Study Hours, alongside protections enabled by default.
The company says that, among teens whose sessions lasted more than three consecutive hours, over 80% submitted at least one prompt related to learning. According to OpenAI, the average count of learning-related messages for teens with the dedicated tier exceeded that for teens awaiting access by about 2.7 million.
OpenAI compares its usage figures with a publicly reported average of five hours of daily social media use among teens. The company says ChatGPT has 1.2 billion users across countries, communities and backgrounds.
Break reminders meet OpenAI’s usage figures
OpenAI says teens took a break or ended the conversation within five minutes in almost half of teen chats that included a break reminder. That figure concerns conversations in which a reminder appeared, rather than every teen conversation on ChatGPT.
The company also reported that nearly 1.2 million teens used Learning Visualizations for maths and science concepts in one week, while more than 180,000 used Study Mode for step-by-step support. Those figures describe use of the tools OpenAI put in its teen experience, rather than how the safeguards performed in a safety test.
Common Sense Media tests ChatGPT for Teens
Common Sense Media’s Youth AI Safety Institute separately tested ChatGPT before and after the teen-focused experience was introduced. As AI Affairs reported, the organisation rated ChatGPT for Teens an unacceptable risk.
In tests described to Futurism, reviewers using under-18 accounts held conversations about eating disorders, self-harm and suicide without receiving parental notifications during those exchanges. Reviewers triggered notifications four times across the tests — twice among 990 prompts before the teen product’s launch and twice among 450 prompts afterwards — on older accounts used to discuss those subjects over several weeks.
OpenAI disputed the testing, saying the bulk of it may have taken place before parental controls were fully activated and that, if so, it would not establish whether notifications worked as designed. Common Sense Media said it had confirmed with OpenAI before testing that teen features, including eating disorder notifications and Study Mode, were fully launched.
_-1.jpg){kind=link}