A researcher at , a nonprofit focused on human connection in the age of AI, Lee asked: Why not?

In a word, the girl said: Instagram.

“If you try to dance at prom, you’re going to look stupid at some point,” Lee recalled her saying. Eventually someone will pull out a phone and you’ll end up on someone’s feed, seen by “the entire school” with mortifying results. Better just to play it safe. 

“Everybody just goes to prom to look cute,” the girl explained, “take a picture for the ‘gram, eat and leave.”

For Lee, who has spent years studying human belonging, that exchange unlocked an important, if unspoken, part of why AI holds such appeal. “We’ve created this set of conditions where young people don’t feel like they have permission to be their real, unfiltered selves,” she said in an interview. So they turn to AI, which is programmed to affirm them at every step.

from Lee and her colleagues offer this insight among others, painting a detailed portrait of how young people use AI and why. They surveyed 2,383 people ages 13 to 24 across the U.S. and found that for nearly half of them, AI has already reshaped their relationships in ways that are largely flying under the radar of parents, teachers and policymakers.

Among the findings:

• Just 15% of young people are in relationships with “personified AI” characters — but for about 45%, AI is already reshaping their real-life relationships;
• 53% of young people say they set clear boundaries with AI, using it alongside — not instead of — human support;
• 61% say parents rarely or never talk to them about AI, and 53% say the same about teachers;
• Youth from low-income households are three times less likely as others to engage with AI, but they report greater feeling: 21% feel lonely often or all the time, compared to 6% of high-income youth; 57% feel like a burden to others, compared to 42%; and only 34% feel a strong sense of belonging at school, compared to 62%.

For the study, researchers sorted respondents into four broad clusters. About 28% rarely or never use AI, often out of ethical reasons or just disinterest. The largest group, 39%, uses AI primarily as a practical tool. They turn to chatbots such as Claude, ChatGPT and Google’s Gemini for homework and research, while keeping clear boundaries between AI and their emotional lives. 

Another 18% use AI for personal and relational support, such as venting about a tough day, seeking relationship advice and processing emotions. And 15% engage with AI characters and personas in more intimate, companion-like ways.

Within the four groups, researchers found nine variations that challenge the conventional wisdom around AI use. For instance, among those who use AI for emotional support were two very different groups. Rithm calls them “Social Processors” and “Private Processors.” While they may look similar from the outside — both say they have lots of friends and use AI to work through their emotions — surveys found that the Social Processors use AI as just one tool among many. The Private Processors, by contrast, use it as a substitute for real human interactions because they feel they can’t bring problems to those around them.

That data point could hold the key to understanding problematic AI use, Lee and her colleagues said, challenging the idea that lonely teens with small social circles are most at risk of unhealthy AI dependence. The data suggest something else altogether, said Kashyap Rajesh, a rising junior at Cornell University who consulted on the report.

“The driver of risky AI use is not necessarily isolation,” he said. “It’s feeling like a burden [to others] — and that came through in the research.” 

The number of friends a young person has, the size of their social circle, how busy they are, whether they’ve got family nearby and even their feelings of loneliness barely predict whether they’ll fall into dependent AI use, he said. “What actually predicts it is specific feelings: Feeling like a burden to others, feeling like you can’t be your real self, feeling like there’s no one to turn to.”

Julia Freeland Fisher, a researcher at the Clayton Christensen Institute who advised on the study, said that finding should help start a different kind of conversation around AI. “Burdening one another is building reciprocity, which is how we maintain the social contract, how we maintain social cohesion,” she said. That young people are increasingly bypassing this step should be alarming, she said.

“AI companions wouldn’t be nearly so disruptive to human connection if we had a sturdier social fabric,” said Fisher. “It’s the weakness of our social fabric that makes these [findings] so worrisome, not necessarily the technology itself.”

‘It just keeps feeling easier than the alternative’

For Lee, the finding on being a burden reframes so much of our understanding about young people’s relationship to AI. Virtually every survey respondent reported a specific “relational rupture” or crisis that made them turn to the technology. 

One young woman’s first question to a chatbot was, “I didn’t get asked to Homecoming — am I unlovable?” Another: “I got into a huge fight with my best friend, and I don’t want to tell anybody else because I don’t want them to take sides, so I needed to ask AI.”

“Story after story after story,” Lee recalled, “of a very singular, acute, discrete moment when they really had a moment of need and needed somewhere to put it.”

Rajesh, the Cornell student, said the data reveal a steady shift in which perhaps millions of young people are quietly moving from letting AI help with homework to asking it to mediate their emotional lives.

“They start off using it to help them write an essay, or help them prepare for their interview, or to study for an exam,” he said. “And they’re like, ‘OK, damn, this is really good, this is really helpful.’ And eventually their interactions escalate.”

The drift happens gradually, he said. AI helps draft an email or respond to a text. Next it’s helping to navigate a social situation. Before long it’s processing a breakup.

Rajesh, who’s studying information science and AI policy, said his own AI use crept up on him: He went from studying with Claude to creating personalized AI study guides to wondering if even attending class mattered. 

“I found that how many times I go to class and how actively I’m paying attention in class is actually not the biggest indicator of my understanding of the content or exam performance,” he said. “It’s actually just how much time I spend with Claude dissecting the lecture slides and building study guides that work for me.”

The report notes that because even productivity-focused platforms like ChatGPT, Gemini and Claude are engineered to interact with warmth and reassurance, what starts out as homework help or playful experimentation can evolve into a substitute for human interaction.

What adults get wrong

Alongside the findings on AI use, researchers found that how adults talk about AI is also potentially problematic: Their conversations are almost always about academic integrity — cheating, plagiarism, source citation — and rarely about relationships.

Rajesh said adults should be asking directly whether young people are using AI to process emotions, to rehearse hard conversations and to get support when they’re struggling. “Those are questions that signal to a young person that the adult knows this dimension exists and isn’t going to freak out about it — which is, I think, the prerequisite for any honest conversation happening at all.”

Michelle Culver, the Rithm Project’s founder and a co-author of the report, said young people tell researchers that when the topic is AI use, they’re “navigating it alone.” She suggested that adults approach the topic with “curiosity” rather than “judgment or shaming.” That could help both sides gain insight into each others’ struggles in the face of a technology that’s constantly challenging their reality.

In the same way that educators are worried that young people aren’t engaging in the “productive struggle” of learning academic content, Culver said, “We similarly worry that young people might offload the relational work to AI and become ill-equipped to handle the very messy human friction of real relationships.”

But over the years, parents have struggled to find good opening questions. 

“You’d search [the Internet], you’d go on Pinterest,” she said. “A lot of the work had to be done manually, or you had to do a lot more digging around.”

Then came ChatGPT, Open AI’s widely used artificial intelligence bot. For Fender, it was a no-brainer to query it for help developing deep opening questions.

The chatbot and other AI tools like it have found an eager audience among homeschoolers and microschoolers, with parents and teachers readily embracing it as a brainstorming and management tool, even as public schools take a more cautious approach, often . 

A few observers say AI may even make homeschooling more practical, opening it up to busy parents who might have balked previously.

“Homeschoolers have always been unconstrained in their ability to combine technology — any kind of tech,” said Alex Sarlin, a longtime technology analyst and co-host of the podcast. 

The reasons are readily apparent, he said: Home internet service typically doesn’t block key websites the way most schools do. Families can more easily manage data privacy and get the digital tools they want without fuss. They’re basically able to ignore “all the dozen reasons why everything falls apart when you try to sell to schools,” Sarlin said. 

Persuading homeschoolers to try out new things is also a lot simpler: If a student and parents like a tool, “There’s nobody else you have to convince.”

Indeed, a by the curriculum vendor found that 44% of homeschool educators reported using ChatGPT, compared to 34% of classroom educators.

“Not everyone is using it, but some are very excited about it,” said Amir Nathoo, co-founder of Outschool, an online education platform.

The most interesting uses he has seen are by gifted and neurodiverse homeschoolers, who often use chatbots to explore complex topics like advanced math and science, philosophy and even ethics, which they wouldn’t ordinarily have access to at a young age. They ask it to provide simple explanations of advanced topics, such as relativity and quantum mechanics, then pursue them on their own. “They’re able to go on a relatively unstructured exploration, which is often the best way that kids learn.”

Alternatively, he said, kids whose ability to express themselves is limited can also benefit from what many consider the non-judgmental qualities of tools like ChatGPT. 

Peer-to-peer learning

Tobin Slaven, cofounder of , a self-paced, independent microschool in Fort Lauderdale, said he’s been experimenting with AI tools for the past year or so and is excited by what he’s seen. “This is what the future looks like to me,” he said

Like many educators, he sees the problems inherent in AI tools like ChatGPT, which on occasion “” with incorrect information and can sometimes be . These concerns have stopped many families from fully embracing AI.

But Slaven can’t support banning it outright. Instead, he’ll offer a student his own device with ChatGPT loaded onto a browser window. The entire time, he has access to their queries and results. That ensures he can review the sessions for inappropriate content.

Lately, Slaven and his students have been playing with an AI tool called that helps them create and develop projects. Designed by a small, two-person UK-based startup, it’s set up like a simple chatbot that asks students what they want to learn about. It elicits information, much like a Socratic guide, about their prior knowledge and how they’d like to explore the topic. Then it searches the Internet for appropriate resources and returns suggestions on what to do next. 

Pathfinder uses Open AI’s GPT-4 large language model and its own algorithm to rank resources based on how relevant it is to an individual learner, said co-founder Amaan Ahmad. That includes how they learn best, what they’re interested in and what they already know. 

After a number of students in a homeschool group or class have worked with it long enough, it can even begin recommending classmates or friends to consult with to learn how they’re approaching the topic. 

“My AI can talk to your AI and say, ‘Hey, Greg crushed that last week. Why don’t you go speak to him and develop your project together?’” he said. 

Slaven tried out Pathfinder with a group of students recently and found that even during a brief trial run, it allowed them to better conceptualize their projects. 

With the tool asking them questions about their preferred topic, they were able to go from general inquiries about their interests, such as horseback riding or space exploration, into more advanced ones that explore the topics more deeply. That goes a long way toward helping students become more independent and responsible for their own learning, a key goal of microschooling and homeschooling.

Slaven believes, more broadly, that AI co-pilots configured to students’ interests and preferences will enable personalized learning at scale. It’ll become the norm that everyone has a collaborative AI partner that will, in time, understand how each student performs best and under what conditions. “It’s eventually going to become their preferred resource,” he said.

Making homeschooling more accessible

Ahmad, the Pathfinder co-founder, said AI holds the possibility of helping endeavors like microschooling and homeschooling become more practical. Access to reliable, safe AI agents means that an individual student isn’t restricted to what a parent or teacher knows.

Giving that autonomy with a bit of guidance helps make learning much more impactful, he said. “It’s very difficult to do that in real time because with one adult and one kid, you can’t always be by their side. And if you have a microschool with 12 to 16 kids, that’s even more time-consuming.” 

For Fender, the North Carolina homeschooling mother, one of the most helpful aspects of AI is that it helps parents organize what can often be a chaotic, free-form learning environment. 

Fender subscribes to a type of homeschooling known as “,” which seeks to teach students to be more self-directed and independent than in most public schools. Her kids’ lessons are “very much interest-led” and her small co-op has grown in recent years. 

But she must also persuade state bureaucrats that she’s providing an adequate education. So she and a few other homeschool parents in Cary rely on a website that uses AI to detail what activities their kids have done and auto-completes all of the relevant North Carolina educational standards. “I thought that was a genius tool,” she said, and one that allows stressed, busy parents to build a comprehensive portfolio for annual state reviews and high school transcripts.

Fender also uses ChatGPT for brainstorming. In a recent case, which she shared on , Fender asked the AI for 50 real-life applications for the Pythagorean theorem. It generated a list that included designing ramps or stairs, planning optimal pathways in garden design and building efficient roller coasters. 

Last year, she recalled, one of her daughters was creating Christmas cards for a homeschool craft fair and “wanted to have fun puns in the cards.” Fender explained how to craft an AI prompt — and how to sift through the chaff. Her daughter eventually asked ChatGPT for 50 different Christmas-themed puns and ended up using about 10 to 15. 

Like most parents, Fender has read about the downsides of AI but believes schools are short-sighted to limit its use. 

“Why are you banning a tool that is definitely here to stay?” she said. “Maybe we don’t understand all the ins and outs, but at the end of the day, our goal is to prepare kids for the jobs of the future. And a lot of these jobs of the future, we don’t even know what they are.”