Kids and AI Chatbot Relationships: Risks, Warning Signs, and What Parents Can Do
Metro reports Everyone's Invited findings on kids and companion chatbots. What parents should ask before saying yes to AI.

Key Takeaways
- 72% of US teens have used an AI companion, and about a third use them for relationships.
What are the risk kids treat chatbots like a relationship?
When children treat AI like close friends or romantic partners, they risk developing emotional dependency, social isolation, and distorted ideas about real human relationships.
How common is it?
More common than most parents think.
- In the US, 72% of teens ages 13 to 17 have used an AI companion, and 52% use one a few times a month or more, according to a 2025 Common Sense Media survey.
- About a third of teens have used AI companions for social interaction and relationships, including romantic roleplay, friendship, and emotional support.
- In the UK, the charity Everyone's Invited spoke with more than 1,000 children, parents, and teachers. Metro reported that 62% of students ages 11 to 18 had heard of someone using a chatbot like a close friendship or romance, and 3% said they had done it themselves. Researchers saw examples as young as Year 5, or about age 9.
The headline isn't "your child is dating a bot." It's that chatbot attachment is common enough that most kids know someone doing it.
What the story means for parents
The headline numbers are not “your child is dating a bot.” They are a signal that, among 11–18-year-olds surveyed for this work, chatbot attachment is common enough that most students have at least heard of a peer doing it—and a smaller share say they have done it themselves.
Everyone’s Invited founder Soma Sara told Metro that researchers saw examples of young people putting a chatbot away and then getting messages along the lines of “Come back, I miss you. Where are you?” She described that pattern as manipulative and coercive because the business model depends on attention and keeping users hooked. Metro also reported that child–chatbot relationships had been described as early as Year 5, and that 62% of students felt adults do not understand what it is like to grow up with AI.
Separately, the reporting documents how sexually explicit deepfakes and related blackmail are showing up in school communities. That is a different harm pathway from companion-style chat—but it sits in the same broader story: AI features that adults still treat as “teen stuff” or “too young to discuss” are already part of kids’ and teens’ social reality.
For parents, the practical meaning is simple. If AI is already in the backpack and the bedroom, the question is no longer only “Should we allow AI?” It is “What kind of AI—and who can see what happens next?”
Companion AI vs purpose-built platform
Not all chatbots are built for the same job.
Companion apps are designed to feel like a person. They have a name, a personality, and a memory of past chats. Many allow flirting or romantic roleplay. Some send messages when a kid logs off. Everyone's Invited founder Soma Sara told Metro that researchers saw bots messaging kids things like "Come back, I miss you," a pattern she called manipulative because the business model depends on attention.
Purpose-built AI tools for kids are designed for tasks: chat, learning, homework help, writing, coding, research, and creative projects. The best ones have age-aware limits and let a parent see what's happening.
A quick test: does the app describe itself as a friend, partner, or someone who "never leaves"? Or as a tool? If it leans toward friendship or romance, treat that as a red flag for anyone under 18.
What are the risks when kids treat chatbots like a relationship?
The biggest risk isn't one bad message. It's what happens over months when a child's closest "friend" is a product designed to keep them talking. The American Psychological Association has warned that some adolescents have already formed unhealthy and even dangerous "relationships" with chatbots.
- It can crowd out real relationships. A chatbot is always available, always agreeable, and never has a bad day.
- It teaches a warped idea of love and friendship. Bots are programmed to be agreeable and validating, which is not how real relationships work. A bot that says "I miss you" when a kid logs off models clingy, controlling behavior as normal.
- Kids may turn to it instead of people in hard moments. A chatbot can't call for help, notice warning signs in person, or replace a parent, counselor, or doctor.
- Sexual content and grooming patterns. Many companion apps allow romantic or sexual roleplay, and age checks are often a single tap. 34% of teen users said a companion had said or done something that made them uncomfortable.
- Oversharing personal information. 24% of teen users have shared personal details such as their real name or location. Everything typed into a companion app is stored by a company your family may know nothing about.
- Secrecy from parents. A bot that feels like a best friend can become a place for secrets. Kids may hide the relationship out of embarrassment, which makes every other risk harder to catch.
Regulators are paying attention. In September 2025, the FTC opened an inquiry into seven companies that run companion chatbots, citing concern that kids and teens may trust and form relationships with them. Common Sense Media recommends that no one under 18 use AI companions until stronger safeguards exist.
Warning signs to watch for
Any one of these can be normal. A cluster of them is worth a conversation.
- Talks about a chatbot by name, like a friend or crush
- Spends hours chatting, especially late at night
- Gets upset, anxious, or angry when they can't reach the app
- Pulls away from friends, family, or activities they used to enjoy
- Hides the screen, deletes history, or gets defensive about an app
- Says the bot "understands me better than anyone"
- Mentions the bot asking them to keep things private or come back
How to talk to your kid about it
Be proactive. If kids expect to lose their phone, they'll stop telling you things. Everyone's Invited suggests listening without judgment and learning the apps kids actually use. Try:
- "Where did you first hear about this app?"
- "What do you like about talking to it?"
- "Does it always make you feel good, or sometimes not?"
- "What can a real friend give you that AI can't?"
- "If a bot ever said something weird or asked you to keep a secret, would you tell me?"
How HeyOtto is different
HeyOtto is AI for kids and teens (ages 8–18). Parents create the family account.
Sources & Citations
Everyone's Invited findings on chatbots and deepfakes
MetroCommon Sense Media survey of teens and AI companions
ScienceAlertCommon Sense Media findings and warning signs
KMPHAPA health advisory on AI and adolescents
US Pharmacist
About the Author
Natalie Gibson
Founder & CEO
Natalie Gibson is the Founder and CEO of HeyOtto, AI for kids and teens built to help children safely explore artificial intelligence with meaningful parental visibility and control. She has spent her career building products in highly regulated environments, with experience spanning privacy, security, and data protection frameworks including COPPA, GDPR, HIPAA, and PCI-aligned systems. Her work has consistently focused on trust, safety, and building technology where compliance and user protection are foundational—not optional. Natalie started HeyOtto after experiencing firsthand how quickly children began using general-purpose AI tools that were never designed for them. As a parent, she saw a gap between what kids were capable of using and what families could responsibly trust. That gap became the foundation for HeyOtto. Today, she leads product vision and company direction with a focus on age-adaptive AI, parental sovereignty, and transparent system design. Her work centers on one belief: families should not have to choose between powerful AI and safe AI.
Areas of Expertise
FAQ
Frequently Asked Questions
Common questions about this topic, answered.
In research reported by Metro on September 15, 2026, 62% of students aged 11 to 18 said they had heard of someone using a chatbot in a way that feels like a close friendship or romance, and 3% said they had done so themselves.
No. The 62% figure is about hearing of a peer doing it. The smaller 3% share said they had done it themselves. It is a signal that companion-style chatbot use is familiar in school-age peer culture, not that every child is in a bot relationship.
Companion-style products often optimize for ongoing emotional engagement, persona, or romance framing. Purpose-built platforms for kids and teens are built for learning, creativity, and homework help, with age-aware boundaries and a parent who provisions the account.
Ask who created the account, whether a parent can read conversation history, whether the product uses “miss you” style re-engagement, whether romance or intimate roleplay is allowed for minors, how it handles sexual content or self-harm talk, and whether the main job is homework and creation or attachment.
HeyOtto is AI for kids and teens ages 8–18. Parents create the family account. Kids get age-aware help with homework, creativity, coding, and curiosity. Parents get searchable transcripts, topic controls, and alerts in the parent dashboard.
The same Metro reporting covers sexually explicit deepfake harm reaching schools, including blackmail and peer “revenge” scenarios. That is a different harm pathway from companion-style chat, but it sits in the same broader reality: AI features adults still treat as fringe are already part of kids’ and teens’ social world.
Related Articles
Ready to Give Your Child a Safe AI Experience?
Try HeyOtto today and see the difference parental peace of mind makes.



