We Need Rules and Ratings to Maximize the Potential of Generative AI for Kids

Over the last few months, generative AI technologies have exploded in the public consciousness. The emergence of ChatGPT has raised questions about the future of this technology and its potential for both opportunity and harm, especially for kids, as more platforms and companies jump into the fray.

Unfortunately, we find ourselves at a point where the opportunities remain in the realm of possibility—while the harms are present and very real. For 20 years, we at Common Sense have been helping families make great media and technology choices. And in that time we've also called on government and industry leaders to make the digital world safer and healthier for kids. In keeping with that commitment, we have been exploring the potentials and perils of generative AI, and asking parents and kids what they really think about the technology. Our new position paper and poll revealed a few things:

Parents have a lot to learn about AI to catch up to their kids, and kids are using it without their parents' or teachers' knowledge.
From usage to knowledge of ChatGPT, parents lag behind where students age 12–18 already are. Only 30% of surveyed parents said they have used ChatGPT or know a lot about it, compared to a majority of students. This gap matters because kids are already using these tools. A majority of kids surveyed (50%) said they have used ChatGPT for school, and a larger majority (56%) have a friend or classmate who has done so. Half of students have used ChatGPT for school assignments, and of those, nearly all have used it once a week or more. That's significantly higher use than their parents at home. And their use of ChatGPT is mostly sanctioned by their teachers: 48% of students said their teachers mostly allow them to use ChatGPT.

While parents and students are still assessing the new programs, both believe that AI tools will help kids learn.
Generative AI could bring many opportunities to education, from more deeply personalized learning and greater access to high-quality education, to increased creativity, engagement, motivation, and support for teachers. Our poll revealed that 77% of parents and 92% of students are interested in AI-powered education tools that could help students learn.

The current landscape of AI products and platforms don't have meaningful guardrails in place to stop harms from occurring.
The harms of generative AI are very real, including the proliferation of misinformation and biased content, privacy and data violations, and the potential for cheating, among others.

And both parents and students recognize those harms: 60% of parents and students surveyed said that the use of ChatGPT to cheat in school is a major concern, while two-thirds of parents and about half of students also expressed significant concern that kids could become too dependent on ChatGPT for school assignments.

Parents and educators are looking for guidance and guardrails that not only make generative AI safer for kids, but also help schools and teachers use the technologies to improve learning outcomes.
Nearly half of the students we surveyed said their school has policies outlining how ChatGPT can be used (46%), but it's important to note that nearly 1 in 5 (19%) students said their teachers have never mentioned ChatGPT at all—and an identical percentage of respondents said their teachers didn't allow the use of ChatGPT in their classrooms. That means the technology isn't fully on everyone's radar yet, and that there's time to start making substantive changes and regulations. Our survey shows that 61% of parents and 51% of students think schools should limit the use of programs like ChatGPT until rules and safeguards can be put in place.

Our report also reveals that a majority (56%) of parents surveyed don't feel there is a reliable source of information about generative AI, and 82% of parents agree they want a rating system from a third party to help them evaluate AI programs.

In response, Common Sense plans to lead the charge in rating and reviewing generative AI tools as part of our industry-leading media reviews. As always, we will focus on helping parents and educators assess the safety and appropriateness of these tools. In the meantime, caregivers need to ask critical questions about generative AI tools before using them with children. Our resources for educators, along with our privacy review of ChatGPT, are places for parents to start.

Like smartphones and social media 20 years ago, generative AI will fundamentally change the very fabric of how kids experience the world. We will continue to be a voice for thoughtful regulation and transparency in AI, and will do continued research and work around kids and AI in the coming years. As we monitor where this latest wave of technological innovation takes us, we believe that we—collectively, as parents, caregivers, educators, and policymakers—can come together to make sure that generative AI tools are built with kids in mind.

METHODOLOGY
Impact Research, on behalf of Common Sense Media, conducted a nationwide online poll of N=1,181 parents with children age 5–18 enrolled in grades K–12, and N=300 students age 12–18 enrolled in grades K–12. Interviews were conducted March 28–April 9, 2023 and were conducted in English and Spanish. If parents had more than one child, they were asked to answer about their oldest child enrolled in grades K–12. The 95% confidence intervals are ± 2.9 percentage points for the parent results and ± 5.7 percentage points for the student results. The confidence interval for subgroups varies and is higher.