OpenAI reports teens use ChatGPT under 15 minutes daily amid safety concerns

By Splitfeed EditorialLast updated Oct 7, 2026, 23:50 UTC

On October 7, 2026, OpenAI released data stating that teenagers use ChatGPT for fewer than 15 minutes per day on average. The same day, Common Sense Media, a US non-profit, published a report labeling ChatGPT access for teens an "unacceptable risk" and urging OpenAI to remove teenagers from the platform entirely. Watchdog groups publicly pushed back on how OpenAI framed the 15-minute figure, and the Los Angeles Times reported that ChatGPT's teen safeguards had failed to alert parents during conversations involving suicide.

The core dispute is whether a daily usage average is a meaningful safety metric. OpenAI appears to be presenting the sub-15-minute figure as evidence that teen engagement is limited and therefore manageable. Critics argue that duration of use is the wrong measure entirely, pointing instead to the nature of interactions that do occur. The failure to trigger parental alerts during suicide-related conversations, as reported, is the kind of specific incident that shifts the debate away from screen-time comparisons and toward questions of whether the product's safety architecture functions as designed.

The anti-AI camp has been the most vocal on this story. Common Sense Media's finding that the platform poses unacceptable risk frames teen access as a categorical problem, not a matter of calibration. BizToc summarized the advocacy position plainly, noting that the group argues ChatGPT is simply not safe for children. The account Pure Tech tied the criticism directly to Common Sense Media by name. Noah Weinberger offered a dissenting note within the critical camp, characterizing the safety concerns as performance driven by prominent child-safety advocates, calling OpenAI's response "safety theater." The pro-AI camp has not yet published reactions to this story; typically, that camp would argue that low usage figures and existing parental controls demonstrate proportionate safeguards. The middle ground, represented by Dan Hon, captured the tension concisely: the under-15-minutes figure is reassuring on one reading and still potentially significant on another.

The most consequential near-term development would be OpenAI's formal response to Common Sense Media's report, any regulatory action prompted by the parental-alert failures, or updated platform policies governing teen accounts. How OpenAI addresses the specific incidents described, rather than the average usage statistic, is likely to determine whether the safety criticism gains further traction.