Errata Carmona
For parents who expect their children to grow up, live, and work in a world saturated by artificial intelligence, helping their kids become fluent with the technology may feel imperative. And yet, many of these same parents may have reservations about exposing their children to AI. Among other concerns, some research has found that unrestricted AI use among students is associated with diminished cognitive skills.
So how are parents balancing the desire to prepare their kids for an AI world with their worries about AI’s potentially detrimental long-term effects?
Research from a team that includes Chicago Booth’s Alex Imas finds that parents’ fear that their children will fall behind can overshadow concerns about these potential adverse effects. And much of that fear is driven by how much other people’s children are using AI.
In order to gauge parental attitudes and decision-making on the use of AI, the researchers recruited a sample of approximately 2,000 parents of children aged 13–18 across the United States, Canada, and the United Kingdom through the online platform Prolific. They conducted a series of experiments to understand what affects how much parents value AI access for their children, as reflected by their willingness to pay for a three-month premium AI subscription to use for homework help, grammar assistance, and research.
The researchers divided the participants into two groups. A control group was given positive information on the effects of AI on children’s learning: “Recent research has shown that using advanced AI models helps improve performance on writing and language tasks while it is being used.” The treatment group received the same information with one main addition: “. . . A field experiment with high-school students found that access to GPT for learning led to a nearly 20% decline in quantitative reasoning scores when students were later tested without AI.”
Participants then considered a series of scenarios in which AI adoption by teens in their region was 20, 40, 60, or 80 percent. For each scenario, they indicated how much they would pay for the AI subscription.
The findings suggest that social pressure tends to play a key role in parents’ choices regarding their teens’ use of generative AI tools for school and other work. Demand for the AI subscription increased with the percentage of teens in the region using AI, with willingness to pay rising by $1.83 for each 10 percentage point increase in peer use. Compared with what parents were willing to pay when considering the 20 percent adoption scenario, willingness to pay in the 80 percent scenario was more than 60 percent higher.
The pressure to keep up
Parents were willing to pay more for a premium AI subscription as AI use among their teenager’s peers increased. This willingness barely changed even after learning about the technology’s potential harms and not just the benefits.
Parents who received the negative information about AI use were far more likely to believe the technology could have detrimental cognitive effects, and to support an outright ban on AI use in schools—but were no less willing to pay for an AI subscription for their kids.
Why? Twenty percent of parents who supported a ban but also permitted AI use reported fear that their kids might fall behind, while more than 70 percent of parents said AI use would help their kids in the short run, even though it might have negative long-term effects. Overall, parents influenced by reports of teens’ AI use didn’t believe AI was beneficial so much as socially necessary.
The researchers term this social pressure “rat-race dynamics” and argue that simple information campaigns directed at parents that communicate the risks of AI use may not be effective in changing behavior. Instead, the researchers write, “collective action, such as school-level adoption of structured AI tools or coordinated adoption decisions, may be required to achieve socially optimal outcomes.”
The researchers note that rat-race dynamics affect more than just education. For instance, companies may overinvest in AI to avoid falling behind, even if it isn’t ultimately profitable. Awareness of this, they write, is crucial to anyone trying to negotiate the complex trade-offs posed by technological transformation.
Leonardo Bursztyn, Alex Imas, Rafael Jiménez-Durán, Aaron Leonard, and Christopher Roth, “Social Dynamics of AI Adoption,” Working paper, November 2025.
Your Privacy
We want to demonstrate our commitment to your privacy. Please review Chicago Booth's privacy notice, which provides information explaining how and why we collect particular information when you visit our website.