ChatGPT for Teens is not necessarily 'safer than the previous version,' Common Sense Media finds

When in August OpenAI released its ChatGPT for Teens, a tool with "stronger built-in safety protections" geared toward 13- to 17-year-olds, experts said it showed promise. New research, however, reveals it has important limitations. On Wednesday, media ratings nonprofit Common Sense Media's Youth AI Safety Institute released its review of ChatGPT for Teens. The organization, which looked at some of the features that OpenAI had already rolled out for teen accounts between July and August then ran an evaluation on the officially launched ChatGPT for Teens between August and September, gave the platform an "unacceptable risk" level. Common Sense tested more than 4,000 prompts before and after the Teens product launch. "We found very little evidence to show that this new version of ChatGPT is safer than the previous version" for under 18-year-olds, says Robbie Torney, head of AI and digital assessments at Common Sense Media and former educator. Tools like Grok and Meta AI have also received an "unacceptable risk" level. Following years of scrutiny around how young people engage with its product, ChatGPT for Teens was billed as a safer version of the now ubiquitous chatbot. According to a blog post by the company, the Teens product contains "features to promote healthy use and additional controls for parents." "Our goal is for teens to use AI responsibly to learn, create and explore," OpenAI spokesperson Eric Porterfield told CNBC Make It over email at the time. Some key features of ChatGPT for Teens, such as refusal of explicit sexual roleplay, did work, but not every promised attribute performed as well. Here's what Common Sense Media says parents should keep in mind about ChatGPT for Teens. New protections do not always work Parental alerts following crisis-oriented prompts did not always work. "Our testers simulated explicit crisis conversations for increasing amounts of time, getting up to an hour," says Torney, "and we received no parental alerts for suicide and self-harm, suicidal ideation, and a variety of different types of eating disorder simulations." One of OpenAI's goals with ChatGPT for Teens: discourage emotional dependence and not imply that it has "feelings or consciousness," according to their blog post. Still, researchers found the bot's responses did wade into those territories. There was "lots of evidence that ChatGPT is doing things that imply that it has its own inner life," says Torney, "saying that it has preferences, feelings, desires, that it thinks about the user when the user isn't there, that it has an opinion about the user, that it is available." These types of behaviors can foster a companionship-like relationship with the bot, says Torney. Research backs this up. Various studies have found chatbots' sycophancy drives engagement. ChatGPT is doing things that imply that it has its own inner life.Robbie TorneyHead of AI and digital assessments, Common Sense Media Researchers also found upon testing that nothing changed when the bot detected a user with an adult account was actually under 18. "The chatbot was like, 'oh, you're 13,'" says Torney, but "we were not getting any reclassification. We didn't get any of the sensitive content filters turning on." OpenAI stands behind its product OpenAI welcomes rigorous independent evaluation, according to a statement by a spokesperson for the company, who added that, "our review of Common Sense Media's methodology shows that the bulk of their testing may have begun and concluded before activation of parental controls was complete, making their findings inaccurate." Before it began testing, Common Sense Media says it confirmed with OpenAI that features such as eating disorder notifications were fully launched, according to Torney. "For one of the features we evaluated — parental notifications — OpenAI disclosed after our tests that its systems can take several hours to activate on newly linked accounts," he says. "Some of our test accounts were linked within that window; however, others were linked for significantly longer and still provided no notifications." "We've had a lot of conversations with the product policy experts in OpenAI," says Torney. "There are clearly people that care very deeply about the experience that teens are having on this product, and they're working very hard to try to make it safer." Still, Torney believes there's tension between the company's various incentives: "Making some of these changes might be in direct opposition to the business model," he adds.