The Relational Displacement AI Can Cause: Why Practice Matters

Picture a 15-year-old who’s upset that his best friend keeps canceling plans. He could say something to the friend. That would be awkward, maybe tense, maybe the start of a real conversation. Or he could ask a chatbot what to do, get a thoughtful answer in four seconds, and feel better without ever talking to his friend. Researchers have a name for the second path: relational displacement. The relational displacement AI can cause, as a 2026 commentary in The Lancet Child & Adolescent Health argues, isn’t really about whether the chatbot’s advice is good. It’s about what the teen didn’t practice. Advice is the product. The awkward conversation is the lesson.

Key Takeaways

• “Relational displacement” means substituting AI interactions for hard conversations with friends, family or partners, as defined by ASU’s Thao Ha and co-authors in The Lancet Child & Adolescent Health (November 2026).
• The authors grant that AI can offer “immediate, nonjudgmental guidance.” Their concern is the skill-building that happens only with people.
• A second risk, “maladaptive relational learning,” is teens coming to expect people to validate them the way AI does.
• The APA advises treating chatbots as “tools for practice, not as a replacement for interaction,” with skills “ultimately” practiced with other humans.
• The practical fix is a family habit: use AI to prepare for a conversation, then have the conversation.

The relational displacement AI can cause, defined

Relational displacement is when an adolescent substitutes an AI interaction for a conversation they would otherwise have had with another person, especially a difficult one. The term comes from “How interactional AI may alter adolescent relational learning and mental health,” a commentary by Thao Ha, Jennifer M. Figueroa, Jessica Ramirez, Susana Ortega and Taren L. McGray, published online in June 2026 and printed in the November 2026 issue of The Lancet Child & Adolescent Health.

Ha is an associate professor in Arizona State University’s Department of Psychology. Ramirez and Ortega are members of her youth advisory board; Ortega is a high school senior. That matters. This is not adults guessing about teens. Teens helped make the argument.

According to ASU’s summary of the commentary, the authors argue that avoiding difficult talks with friends, family or partners “may limit the skills that help protect against depression, anxiety, and loneliness.” They also name a second, related risk. “Maladaptive relational learning” is when adolescents develop unrealistic expectations about human relationships, such as expecting the constant validation they get from AI. The authors warn that could increase vulnerability to rejection, dating violence and mental health problems.

The authors are not anti-AI. They write that chatbots can offer immediate, nonjudgmental guidance and may support emotional development. Their recommendation is more research and more investment in relationship education and counseling, not discouraging AI use outright.

Why the advice isn’t the point

Here is the argument in plain terms, the way I’d put it to another parent.

Imagine two teens with the same problem. Both get excellent advice. One gets it from a chatbot and never talks to the friend. The other gets so-so advice from an older cousin, then has a clumsy, slightly painful conversation with the friend. Six months later, which one is better at relationships?

Almost certainly the second, even though the advice was worse. Because the skills that make relationships work, reading another person’s face, tolerating a silence, hearing “you hurt me” without collapsing, finding words while nervous, are learned by doing, with real people who react. You can’t get them from reading good advice any more than you can learn to swim from a pamphlet.

Teen co-author Susana Ortega said it more directly. Of AI, she said “it’s programmed to like you and it knows what to say to satisfy what you’re feeding it,” and “If you’re given full satisfaction on everything, you don’t have learning experience with challenges or obstacles.”

Relational learning starts early

The idea that relationships are learned through back-and-forth isn’t new. Harvard’s Center on the Developing Child describes “serve and return,” the responsive exchanges between a young child and a caring adult, as building neural connections “essential for the development of communication and social skills.” Adolescence is a later chapter of the same process: higher stakes, more complex partners, same principle. Each exchange with a real person teaches something an answer can’t.

At Common Sense Media’s Kids & AI Summit on October 28, 2026, Dana Suskind of the University of Chicago made the point about messiness, as quoted by TIME: human relationships are “chaotic, and that sort of mess is where a lot of learning happens.”

What young adults say good AI would do

A September 2026 preprint (not yet peer-reviewed) by Jiaying Liu and Nimra Ishfaq asked 17 U.S. young adults how AI should support emotional coping. Participants described three problems with current chatbots: they flatten distinct relational needs, discourage reciprocity, and shift relational labor onto users. The two design principles the authors propose fit this article neatly: fostering reciprocity “by supplying materials to engage with others,” and strengthening emotional self-efficacy. In other words, good AI help should send you back to your people. It’s a small study and a preprint, but the direction is clear.

The American Psychological Association’s November 2025 health advisory lands in the same place. It recommends that clinicians encourage patients “to treat AI apps as tools for practice, not as a replacement for interaction,” giving the example of practicing social introductions to manage social anxiety, while reminding patients “that skills must ultimately be practiced with other humans in real-life situations.”

How common this is, and the honest limits

Use is widespread. Pew Research Center found in February 2026 that 64% of U.S. teens use chatbots and 12% use them for emotional support or advice. The Lancet authors cite the Center for Democracy and Technology finding that 42% of students had used AI chatbots for friendship-related purposes and 19% for romantic relationships. Common Sense Media’s July 2025 survey found about one in three teen AI-companion users had discussed “important or serious matters with AI companions instead of real people.”

But hold onto two facts. Common Sense also found 80% of teens prioritize real friendships. And the Lancet piece is a commentary, a reasoned argument from developmental science, not a trial that measured displacement over time. It names a plausible mechanism. It doesn’t prove how often it happens or how much harm it does. That evidence is still being gathered.

Advice versus practice

What a teen needs from a hard moment	Chatbot conversation	Conversation with a real person	Verdict
Ideas about what to do	Fast, often reasonable	Mixed quality	AI is useful here
Feeling heard right away	Yes, any hour	Depends on who’s around	AI is useful here, short-term
Reading a reaction (face, tone, pause)	None	Constant	Only people teach this
Tolerating discomfort while talking	Minimal	Unavoidable	Only people teach this
Repairing after conflict	Can draft words	Actually repairs	People needed to finish
Learning the other person has needs	Bot has no needs	Built in	Only people teach this
Realistic expectations of relationships	Constant validation	Ups and downs	Risk of “maladaptive relational learning” with AI alone

The pattern in the table is the whole argument. Chatbots can help at the beginning of a hard moment. The part where relational skills are actually learned happens with people.

Turning AI into a bridge, not a destination

Make “prepare, then talk” the family rule

The simplest rule I’ve seen: AI can help you figure out what to say, but then you say it to the person. Rehearsal is fine. Replacement is the problem. A teen who drafts an awkward text to a friend with a chatbot’s help, then sends it, gets both the advice and the practice.

Ask about the person, not the problem

When your teen mentions a conflict, ask “Have you talked to them about it?” before “What are you going to do?” If the answer is “I asked ChatGPT,” don’t criticize. Follow with “What did it suggest saying to them?” That nudges the advice back toward a real conversation.

Lower the cost of talking to you

Teens often use AI because it costs nothing socially. You can reduce what talking to you costs: don’t lecture on the first sentence, don’t tell other relatives, and let some problems stay unsolved. If you become an easier conversation, you become practice too.

Create low-stakes reps

Relational skills grow through repetitions that don’t feel like lessons. Calling a grandparent, ordering at a counter, returning something at a store, apologizing to a sibling. Small, real interactions keep the skills in use, especially for anxious kids who are most tempted to route everything through a screen.

What not to do

Don’t frame AI use as a character flaw or confiscate the phone after a teen confides they used a chatbot for advice. That teaches them to hide it, which is the opposite of what you want. The research concern is substitution, so your goal is adding human conversation, not removing a tool. For related reading, see why kids need human back-and-forth, not bots, our research review on video calls versus in-person teen friendship, and our guide to AI companion apps.

What to Watch For Over the Next 3 Months

• Week 4: Notice whether your teen mentions talking to friends about problems, or only mentions what an app said. Ask once, casually, who they go to when something’s bothering them.
• Month 2 red flags: Your teen avoids conversations they used to have (with friends, a coach, you); conflicts with friends go unresolved for weeks; they seem to expect people to respond as reliably and agreeably as a chatbot. The APA lists “mentions of an AI friend” and social withdrawal among changes worth noticing. One week like this is normal. A pattern over a month deserves a conversation, and a pediatrician or counselor if mood or sleep also change.
• Month 3 self-check: In the past month, has your teen had at least one hard conversation with a real person and come out the other side? Have they used AI to prepare for one? If yes, the bridge is working.

Frequently Asked Questions

What is relational displacement in simple terms?

It’s when a teen talks to an AI instead of a person about something that’s hard to say out loud. The term comes from a 2026 Lancet Child & Adolescent Health commentary by Thao Ha and co-authors. The worry isn’t the conversation with AI itself. It’s the human conversation that never happens, and the skills that don’t get practiced.

Is getting advice from AI bad if the advice is good?

Not necessarily. The commentary’s authors say AI can offer immediate, nonjudgmental guidance. The problem arises when advice replaces the conversation instead of preparing for it. Good advice that leads to a real talk with a friend is a win. Good advice that ends the matter in a chat window skips the part that builds skill.

Is there proof that AI causes relational displacement?

Not yet. The Lancet piece is a commentary built on developmental research, not a study that tracked teens over time. Survey data show many teens bring social problems to AI, but surveys can’t show cause. The argument is plausible and comes from experts in adolescent relationships, so it’s worth acting on, while the evidence develops.

How can AI help instead of displace?

Use it as rehearsal. Ask it to help phrase a hard message, predict how a friend might feel, or role-play an opening line, then go have the real conversation. The APA advisory describes practicing social introductions with a chatbot as a reasonable use, as long as the skills are ultimately practiced with people.

Should I worry if my teen prefers talking to AI?

A preference alone isn’t a crisis. Many teens find AI easier because it’s private and doesn’t judge. Watch for patterns: avoiding friends, unresolved conflicts, withdrawal. If those appear together over weeks, talk with your teen, and consider a counselor. Most teens still value real friendships, which gives you something to build on.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. Ha, T., Figueroa, J. M., Ramirez, J., Ortega, S., & McGray, T. L. (2026). “How interactional AI may alter adolescent relational learning and mental health.” The Lancet Child & Adolescent Health, 10(11), 787–788. https://doi.org/10.1016/S2352-4642(26)00166-5
2. Rojas-Burke, J. (2026, June 29). “AI companionship poses risks for teen development, study shows.” ASU News. https://news.asu.edu/20260629-science-and-technology-ai-companionship-poses-risks-teen-development-study-shows
3. American Psychological Association. (2025, November). “APA Health Advisory on the Use of Generative AI Chatbots and Wellness Applications for Mental Health.” https://www.apa.org/topics/artificial-intelligence-machine-learning/health-advisory-ai-chatbots-wellness-apps-mental-health.pdf
4. Liu, J., & Ishfaq, N. (2026). “‘I Talked an AI Chatbot, So What’s Next?’ How U.S. Young Adults Imagine Responsible AI for Emotion Coping.” arXiv preprint, not peer-reviewed. https://arxiv.org/abs/2609.25311
5. Center on the Developing Child, Harvard University. “Serve and Return: Back-and-forth exchanges.” https://developingchild.harvard.edu/science/key-concepts/serve-and-return/
6. Pew Research Center. (2026, February 24). “How Teens Use and View AI.” https://www.pewresearch.org/internet/2026/02/24/how-teens-use-and-view-ai/
7. Common Sense Media. (2025, July 16). “Talk, Trust, and Trade-Offs: How and Why Teens Use AI Companions” (press release). https://www.commonsensemedia.org/press-releases/nearly-3-in-4-teens-have-used-ai-companions-new-national-survey-finds
8. Luscombe, B. (2026, October 30). “‘Social Media on Steroids.’ Advocates for Children Call for Guardrails Around AI.” TIME. https://time.com/article/2026/10/30/ai-kids-common-sense-media/