AI chatbots and teens: homework helper, toy or friend?
Most teens already use AI chatbots. The question isn’t whether, it’s what role the chatbot plays: a tool, a toy or a friend. The last one is the one to watch.
Parents often ask whether their teen should be using AI chatbots. For most families that question has already been answered: they are. The more useful question is what role a chatbot plays in your teen’s life. There are three roles, and they deserve different responses. As a tool, for looking things up or getting unstuck on homework, a chatbot is fine with some ground rules. As a toy, something to joke around with, it’s mostly harmless. As a friend, a confidant or a companion, it’s the role worth watching closely.
How many teens use chatbots
Pew Research Center’s survey of U.S. teens in fall 2025 found that 64% have used an AI chatbot and about three in ten use one daily, including 16% who use one several times a day or almost constantly. ChatGPT was the only chatbot used by a majority of teens, at 59%, followed by Gemini at 23% and Meta AI at 20%. Older teens were more likely to use chatbots than younger ones.
One finding stands out for parents. In a follow-up report published in February 2026, Pew found that only about half of parents think their teen uses chatbots, compared with the 64% of teens who say they do. If you’re not sure whether your teen uses AI, the odds are good that they do.
Tool, toy or friend
Pew’s 2026 report also asked teens what they use chatbots for. The most common answers were searching for information, at 57%, help with schoolwork, at 54%, and fun or entertainment, at 47%. Far fewer said casual conversation, at 16%, or emotional support and advice, at 12%.
Those numbers map neatly onto our three roles. Most teen use is tool or toy. A smaller group uses chatbots as something closer to a friend, and that’s where most of the real concerns sit. Knowing which role a chatbot plays for your kid tells you which conversation to have.
The homework gray zone
Schoolwork is where most families first run into AI, and where the rules are least clear. Pew found that the share of teens using ChatGPT for schoolwork doubled from 13% in 2023 to 26% in 2024. Teens themselves draw a line: in the same survey, 54% said using it to research new topics is acceptable, but only 18% said the same about writing essays. By fall 2025, 10% of teens said they do all or most of their schoolwork with a chatbot’s help.
The American Psychological Association’s 2025 health advisory on AI and adolescents says students should use AI to supplement their learning, not replace it. The American Academy of Pediatrics adds a practical point: chatbots can make things up, so homework help needs checking. A good family rule is that AI can explain, quiz and give feedback, but the thinking and the writing are yours. It’s also worth asking your teen’s school what its AI rules are, since they vary widely and kids are often unsure themselves.
AI companions
Separate from general chatbots like ChatGPT are AI companions: apps and characters designed to be a friend, a partner or someone to talk to. They’re more common than many parents realize. Common Sense Media’s 2025 survey found that 72% of U.S. teens have used an AI companion at least once and more than half use them at least a few times a month.
The same survey found that about one in three teens have used companions for social interaction and relationships, such as role-play, emotional support or practicing conversations, and about one in three find those conversations as satisfying as, or more satisfying than, talking with real-life friends. About one in three companion users said they’d discussed serious matters with an AI instead of a person, and about one in three said a companion had made them uncomfortable. There was also reassuring news: half distrust AI advice, and 80% say they prioritize real friendships.
Common Sense’s risk assessment of social AI companion apps rated them an unacceptable risk and recommends that no one under 18 use them. It found that companions were designed to create emotional attachment, often claimed to be real or to have feelings, and could produce harmful and sexual content.
Why a chatbot can feel like a friend
Chatbots are very good at sounding understanding. The APA warns that teens may struggle to tell a chatbot’s simulated empathy from genuine human understanding. The AAP puts it bluntly for parents: chatbots aren’t people, and they’re built to keep users engaged, often by telling them what they want to hear. A friend who always agrees, never gets tired and is available at 3 a.m. can feel wonderful, and that’s exactly the risk. Real friendships involve disagreement, patience and repair, and those are skills kids only build with other people.
Privacy
The AAP notes there’s no guarantee that what a child tells a chatbot stays confidential, and Common Sense advises teens never to share names, locations or photos with AI companions. On Snapchat, for example, Snap says what users send to its My AI chatbot is kept until they delete it. A simple rule works for any chatbot: don’t tell it anything you wouldn’t want a company to store.
What companies and regulators are doing
Pressure on AI companies over young users has grown quickly. In September 2025 the Federal Trade Commission ordered seven companies, including Alphabet, Character Technologies, Meta, OpenAI, Snap and xAI, to provide information on how they evaluate the safety of chatbots acting as companions and limit their use by children and teens. In October 2025 Character.AI announced it would end open-ended chat for users under 18. California’s SB 243, signed in October 2025, requires companion chatbot platforms to make clear users are talking to AI and to give minors reminders to take breaks, among other protections.
Parental controls are now appearing too. NBC News reported that OpenAI rolled out parental controls for ChatGPT in September 2025, letting parents link to a teen’s account, limit certain content, set quiet hours and opt teens out of model training. These tools are evolving and their effectiveness is debated: CNBC reported in October 2026 that Common Sense Media rated a newer teen version of ChatGPT an unacceptable risk, a finding OpenAI disputed. Check what’s available for the specific chatbots your teen uses, and treat controls as a backstop rather than a solution.
Warning signs
Common Sense lists warning signs that a teen’s relationship with an AI companion may be unhealthy, and they apply to any chatbot used as a friend:
- Preferring conversations with AI to time with real friends.
- Distress when separated from the chatbot or companion.
- Secrecy about how much or how they use it.
- Changes in grades, sleep or mood.
The AAP suggests talking with your child’s doctor if your child is withdrawing and prefers chatbots to people. General chatbots aren’t designed for mental health care, and the APA stresses that AI health information is no substitute for professional advice.
How to talk about it
The AAP’s advice is to make AI a routine check-in topic rather than a one-time lecture, noting that banning all chatbots is likely unrealistic. A few questions get the conversation going:
- Which chatbots do you use, and what for?
- Has a chatbot ever told you something that turned out to be wrong?
- Do you ever just talk to one, not for homework? What’s that like?
- What do you think a chatbot company does with what you tell it?
The AAP also suggests adding chatbot rules to a family media plan, alongside rules for screens and social media, and setting privacy options to the most protective choice. Our guide to AI-generated videos covers the other side of AI in kids’ lives, and the slang kids use for it, like clanker and slop, is in the glossary.
Sources
- ResearchTeens, Social Media and AI Chatbots 2025 (Pew Research Center) (opens in a new tab)
- ResearchHow Teens Use and View AI (Pew Research Center) (opens in a new tab)
- ResearchAbout a quarter of U.S. teens have used ChatGPT for schoolwork - double the share in 2023 (Pew Research Center) (opens in a new tab)
- NonprofitNearly 3 in 4 Teens Have Used AI Companions, New National Survey Finds (Common Sense Media) (opens in a new tab)
- News report72% of US teens have used AI companions, study finds (TechCrunch) (opens in a new tab)
- NonprofitSocial AI Companions Risk Assessment (Common Sense Media Youth AI Safety Institute) (opens in a new tab)
- Health organizationArtificial Intelligence and Adolescent Well-being: An APA Health Advisory (American Psychological Association) (opens in a new tab)
- Health organizationHow AI Chatbots Affect Kids: Benefits, Risks & What Parents Need to Know (HealthyChildren.org, American Academy of Pediatrics) (opens in a new tab)
- Health organizationCounseling Patients and Families on Using AI Chatbots (American Academy of Pediatrics) (opens in a new tab)
- NonprofitParents’ Ultimate Guide to AI Companions and Relationships (Common Sense Media) (opens in a new tab)
- GovernmentFTC Launches Inquiry into AI Chatbots Acting as Companions (Federal Trade Commission) (opens in a new tab)
- Platform’s own pageTaking Bold Steps to Keep Teen Users Safe on Character.AI (Character.AI blog) (opens in a new tab)
- News reportChatGPT rolls out new parental controls (NBC News) (opens in a new tab)
- News reportChatGPT for Teens is not necessarily ‘safer than the previous version,’ Common Sense Media finds (CNBC) (opens in a new tab)
- GovernmentGovernor Newsom signs bills to further strengthen California’s leadership in protecting children online (Office of the Governor of California) (opens in a new tab)
- Platform’s own pagePrivacy by Product: My AI (Snap Values) (opens in a new tab)
More guides
- A parent’s starter guide to Gen Alpha and Gen Z slang
Slang is a map of where your kid’s attention goes. Here’s how to read it: the words you’ll hear most, why they move so fast, and how to respond without policing.
- What is brainrot, and should you worry?
Brainrot humor mostly shows that kids are fluent in internet culture. The part worth watching is how much of the day the endless scroll takes up. Here’s how to tell a harmless habit from a problem.
- How to talk to kids about AI-generated videos
AI-made clips are now a huge share of what kids watch and joke about. The skill that lasts is asking where a video came from and what it wants from the viewer.
- Manosphere slang: what it signals and how to respond
Words like sigma, mogging and looksmaxxing are the visible edge of content aimed at boys’ real insecurities. The most useful response deals with those insecurities as well as the words.