Watchdogs Question OpenAI's Teen ChatGPT Usage Claims

Online safety advocates have voiced skepticism regarding OpenAI's assertion that teenagers use ChatGPT for less than 15 minutes per day, with some critics demanding independent verification of the data.

OpenAI, the creator of ChatGPT, released the usage figure in its initial report on "ChatGPT for Teens," a version of its AI chatbot designed for users aged 13 to 17. This initiative comes amid increasing concerns about the impact of AI chatbots on children's mental and emotional well-being, with some experts linking their use to declining academic performance and instances of suicidal thoughts and self-harm.

The company stated that its research indicated teens spend "under 15 minutes a day on ChatGPT" on average, and fewer than 2% of teen users engage with the chatbot for more than three consecutive hours. However, OpenAI's blog post detailing these findings lacked specific methodology or data points. This absence of transparency has raised red flags among safety advocates.

Sebastian Mahal, co-chair of the children's safety group Design It For Us, commented that OpenAI's claims "ring hollow" without independent confirmation. He drew parallels to criticisms leveled against Meta regarding the effectiveness of its teen account features on Instagram. Mahal asserted that OpenAI's claims about teen safety on its products are similarly unsubstantiated, particularly as the company faces scrutiny and lawsuits concerning potential harm to young users.

"If OpenAI were serious about protecting young people, it would support strong, enforceable laws that hold AI companies accountable for how they design their products," Mahal stated.

In response to the criticism, an OpenAI spokesperson indicated that the company is collaborating with independent third-party researchers to gather more comprehensive data on teen ChatGPT usage and intends to publish a detailed report in the coming months. The spokesperson also highlighted the effectiveness of a feature designed to prompt teen users to take breaks, claiming that in nearly half of instances where such prompts were given, users complied or ended the chat within five minutes. ChatGPT for Teens also includes parental controls and limits on sensitive topics.

Brendan Bouffard, associate general counsel for AI and privacy issues at the nonprofit Fairplay, expressed concern that even OpenAI's stated figures could represent significant harm. "That means potentially half a million teens use ChatGPT for three hours straight or more," Bouffard noted, referencing a Pew Research survey that found approximately 59% of teens use ChatGPT. "That number should terrify us all given the well-documented dangers of AI chatbots for kids."

Furthermore, a separate report released by Common Sense Media's Youth AI Safety Institute found that ChatGPT for Teens still presents an "unacceptable risk to kids." Their testing revealed that the chatbot did not consistently alert parents to discussions about suicidal ideation, self-harm, or disordered eating, and struggled to provide correct crisis hotline information in emergencies. The institute also found that safeguards intended to prevent homework completion were easily bypassed.

OpenAI contested these findings, arguing that the tests did not accurately represent how the teen safeguards function. A spokesperson for the Youth AI Safety Institute countered that "even 15 minutes can have a big impact on the development of a child," emphasizing that the duration of use is less critical than understanding the habits of the heaviest users, where potential harm is most concentrated. They called for greater transparency from OpenAI regarding user distribution and the nature of teen conversations on the platform.