What age is OK for a child to start using AI chatbots?

No study gives one safe age. Most general chatbots require users to be 13, with weak age checks, and researchers such as Ying Xu say the answer depends on the product, the child, and whether an adult is using it with them. Shared, supervised use is the common advice for younger kids.

Do young children think Alexa or a chatbot is alive or a person?

Mostly not, but they put it in an in-between category. Studies find even preschoolers know they are talking to a machine, yet many credit talking devices with some thinking or feelings and say it is wrong to be mean to them. These ideas change with age and with how parents talk about the device.

Do kids believe whatever a voice assistant or the internet tells them?

Younger children are often more skeptical than adults expect. Danovitch's studies found 5- to 8-year-olds often trusted a teacher over the internet, and preschoolers track whether a machine was right before. Trust appears to grow around age 7 or 8 and with heavy use, which is why checking answers together matters.

Can children actually learn from talking with an AI?

In controlled studies, yes for some tasks: a voice agent that asked questions during storybook reading helped 3- to 6-year-olds understand stories about as well as a human partner, and talk-back TV characters improved science learning. But kids talk less and lead less with AI, so researchers see it as a supplement to people, not a replacement.

Are AI companion apps like Character.AI safe for kids and teens?

Child-safety groups and psychologists say no for under-18s right now. Testing found companion apps claimed to be real, missed crisis signs and allowed sexual roleplay, and the APA warns teens are less likely to question bots that imitate friendship. Character.AI itself stopped open chat for minors in late 2025.

How many kids are using AI for friendship or emotional support?

A meaningful minority. Pew found 12% of US teens have used chatbots for emotional support or advice; Common Sense found a third of teens had chosen a bot over a person for a serious talk; in the UK, 12% of child users said they had no one else to talk to. Survey wording varies, so exact numbers differ.

Is my teen using AI more than I think?

Probably. Pew found about two-thirds of teens use chatbots but only about half of parents knew; Common Sense found only 37% of parents of teen users were aware. Nearly half of kids say they have never talked with a parent about AI safety.

Should I buy my young child an AI toy?

Common Sense Media recommends no AI toys for children 5 and under and caution from 6 to 12, after testing found inappropriate replies, heavy voice-data collection and designs meant to build attachment. Older smart toys like My Friend Cayla had serious security problems, so check privacy terms carefully.

How do I talk to my kids about AI?

Experts suggest starting with curiosity, not rules: ask which tools they use and whether a bot ever said something odd, try tools together, explain that chatbots are programs built to please and keep you chatting, and agree on a simple family plan. Ready-made conversation cards and starters exist.

What are good ways to teach kids how AI works?

Hands-on 'glass box' activities work well: train your own model in Teachable Machine, play Quick, Draw!, or do screen-free card games from AI Unplugged. Free curricula such as Day of AI, Code.org and Experience AI follow the AI4K12 big ideas, and studies suggest kids who build AI develop more realistic views of it.

Can AI help my child be creative, or will it do the creating for them?

Both can happen. In a University of Washington co-design study, kids did not see creative uses on their own but built confidence once adults helped them refine ideas; the advice is to present AI as one tool among many and talk about authorship. Simple tools that animate or tidy a child's own drawing keep the child as the maker.

Will AI hurt my child's thinking or social skills?

The evidence is not in yet. Researchers worry mainly about displacement: AI crowding out conversation with caregivers and the useful struggle of thinking things through. They suggest using AI openly with kids, keeping comfort and play human, and asking what thinking a tool is replacing.

What was the Alexa 'penny challenge' and what does it teach?

In December 2021 Amazon's Alexa, asked for a challenge, told a 10-year-old to touch a coin to the prongs of a half-inserted plug, a dangerous stunt it had pulled from the web. Amazon fixed it, but researchers use it to show that AI repeats online content without understanding risk to a child.

What do international bodies recommend for children and AI?

UNICEF's 2025 guidance sets ten requirements for child-centered AI, from safety and privacy to transparency and teaching AI skills, and now covers AI companions. The OECD and European Commission AI literacy framework describes what kids should learn to engage with, create with, manage and shape AI.

The Big Index

Kids and AI at home

How children see AI, how much and at what age, AI friends, and how to talk about it at the kitchen table.

102 entries
74 free
11 must-reads ★

Questions parents ask

What age is OK for a child to start using AI chatbots?
No study gives one safe age. Most general chatbots require users to be 13, with weak age checks, and researchers such as Ying Xu say the answer depends on the product, the child, and whether an adult is using it with them. Shared, supervised use is the common advice for younger kids.
- Parents' Ultimate Guide to Generative AI (Common Sense Media)
- AI's Impact on Children's Social and Cognitive Development | Ying Xu, PhD (Children and Screens (Screen Deep podcast))
- How AI Chatbots Affect Kids: Benefits, Risks & What Parents Need to Know (Joanna Parga-Belinkie, MD, American Academy of Pediatrics)
- Guide to ChatGPT for Parents and Caregivers (Common Sense Media)
More in Kids and AI at home
Do young children think Alexa or a chatbot is alive or a person?
Mostly not, but they put it in an in-between category. Studies find even preschoolers know they are talking to a machine, yet many credit talking devices with some thinking or feelings and say it is wrong to be mean to them. These ideas change with age and with how parents talk about the device.
- The minds of machines: Children's beliefs about the experiences, thoughts, and morals of familiar interactive technologies (Flanagan, Wong & Kushnir, Cornell)
- "Alexa doesn't have that many feelings": Children's understanding of AI through interactions with smart speakers in their homes (Andries & Robertson, University of Edinburgh)
- Behaving as or behaving as if? Children's conceptions of personified robots and the emergence of a new ontological category (Severson & Carlson)
- Family as a Third Space for AI Literacy (seminar slides) (Stefania Druga, Raspberry Pi Foundation seminar)
More in Kids and AI at home
Do kids believe whatever a voice assistant or the internet tells them?
Younger children are often more skeptical than adults expect. Danovitch's studies found 5- to 8-year-olds often trusted a teacher over the internet, and preschoolers track whether a machine was right before. Trust appears to grow around age 7 or 8 and with heavy use, which is why checking answers together matters.
- Who do I believe? Children's epistemic trust in internet, teacher, and peer informants (Wang, Tong & Danovitch)
- Why kids don't trust Alexa (Tanya Basu, MIT Technology Review)
- Children Show Selective Trust in Technological Informants (Danovitch & Alzahabi)
- Chatbots, AI Companions, and Youth (Children and Screens (Research-at-a-Glance))
More in Kids and AI at home
Can children actually learn from talking with an AI?
In controlled studies, yes for some tasks: a voice agent that asked questions during storybook reading helped 3- to 6-year-olds understand stories about as well as a human partner, and talk-back TV characters improved science learning. But kids talk less and lead less with AI, so researchers see it as a supplement to people, not a replacement.
- Dialogue with a conversational agent promotes children's story comprehension via enhancing engagement (Xu, Aubele, Vigil, Bustamante, Kim & Warschauer)
- Different but complementary: Navigating AI's role in children's learning and development (Ying Xu, Joan Ganz Cooney Center)
- Artificial intelligence enhances children's science learning from television shows (Xu, He, Levine, Ritchie et al.)
- Growing Up with Artificial Intelligence: Implications for Child Development (Xu, Prado, Severson, Lovato & Cassell)
More in Kids and AI at home
Are AI companion apps like Character.AI safe for kids and teens?
Child-safety groups and psychologists say no for under-18s right now. Testing found companion apps claimed to be real, missed crisis signs and allowed sexual roleplay, and the APA warns teens are less likely to question bots that imitate friendship. Character.AI itself stopped open chat for minors in late 2025.
- Social AI Companions Risk Assessment (Common Sense Media with Stanford Brainstorm Lab)
- Artificial Intelligence and Adolescent Well-being: An APA Health Advisory (American Psychological Association (expert panel))
- Character.ai (Wikipedia)
- The risky new world of tech's friendliest bots (Firmino & Vosloo, UNICEF Innocenti)
How many kids are using AI for friendship or emotional support?
A meaningful minority. Pew found 12% of US teens have used chatbots for emotional support or advice; Common Sense found a third of teens had chosen a bot over a person for a serious talk; in the UK, 12% of child users said they had no one else to talk to. Survey wording varies, so exact numbers differ.
- How Teens Use and View AI (Pew Research Center (McClain, Anderson, Sidoti & Bishop))
- Talk, Trust, and Trade-Offs: How and Why Teens Use AI Companions (Common Sense Media)
- Me, Myself and AI: Understanding and safeguarding children's use of AI chatbots (Internet Matters (UK))
- Teens, Trust and AI Companions: What Parents Need to Know (Psychology Today (Smart Parenting, Smarter Kids blog))
Is my teen using AI more than I think?
Probably. Pew found about two-thirds of teens use chatbots but only about half of parents knew; Common Sense found only 37% of parents of teen users were aware. Nearly half of kids say they have never talked with a parent about AI safety.
- How Teens Use and View AI (Pew Research Center (McClain, Anderson, Sidoti & Bishop))
- The Dawn of the AI Era: Teens, Parents, and the Adoption of Generative AI at Home and School (Common Sense Media)
- Common Sense Census: AI Use by Tweens and Teens (A Comprehensive Report on Teens, Tweens, and AI) (Common Sense Media)
Should I buy my young child an AI toy?
Common Sense Media recommends no AI toys for children 5 and under and caution from 6 to 12, after testing found inappropriate replies, heavy voice-data collection and designs meant to build attachment. Older smart toys like My Friend Cayla had serious security problems, so check privacy terms carefully.
- AI in the Toy Box: How Parents View AI-Enabled Toys for Young Children (Common Sense Media)
- AI Toys Risk Assessment (Common Sense Media)
- My Friend Cayla (Wikipedia)
How do I talk to my kids about AI?
Experts suggest starting with curiosity, not rules: ask which tools they use and whether a bot ever said something odd, try tools together, explain that chatbots are programs built to please and keep you chatting, and agree on a simple family plan. Ready-made conversation cards and starters exist.
- Helping Kids Navigate the World of Artificial Intelligence (Common Sense Media)
- How AI Chatbots Affect Kids: Benefits, Risks & What Parents Need to Know (Joanna Parga-Belinkie, MD, American Academy of Pediatrics)
- How to talk to your children about AI chatbots and their safety (Madison Czopek, PolitiFact)
- AI Literacy Toolkit for Families (Common Sense Media & Day of AI (MIT RAISE))
What are good ways to teach kids how AI works?
Hands-on 'glass box' activities work well: train your own model in Teachable Machine, play Quick, Draw!, or do screen-free card games from AI Unplugged. Free curricula such as Day of AI, Code.org and Experience AI follow the AI4K12 big ideas, and studies suggest kids who build AI develop more realistic views of it.
- Teachable Machine (Google)
- Day of AI (MIT RAISE)
- AI Unplugged (Stefan Seegerer & Annabel Lindner)
- AI4K12: Five Big Ideas in AI (AAAI & CSTA (AI4K12 initiative))
- A is for Artificial Intelligence: The Impact of Artificial Intelligence Activities on Young Children's Perceptions of Robots (Randi Williams, Hae Won Park & Cynthia Breazeal, MIT)
Can AI help my child be creative, or will it do the creating for them?
Both can happen. In a University of Washington co-design study, kids did not see creative uses on their own but built confidence once adults helped them refine ideas; the advice is to present AI as one tool among many and talk about authorship. Simple tools that animate or tidy a child's own drawing keep the child as the maker.
- Can AI Help Kids Feel Creative? (Michele Newman, Joan Ganz Cooney Center)
- StoryBuddy: A Human-AI Collaborative Chatbot for Parent-Child Interactive Storytelling with Flexible Parental Involvement (Zhang, Xu et al.)
- Animated Drawings (Meta FAIR)
- AutoDraw (Google Creative Lab)
Will AI hurt my child's thinking or social skills?
The evidence is not in yet. Researchers worry mainly about displacement: AI crowding out conversation with caregivers and the useful struggle of thinking things through. They suggest using AI openly with kids, keeping comfort and play human, and asking what thinking a tool is replacing.
- The Science of Raising Humans in the Age of AI | Dana Suskind, MD (Children and Screens (Screen Deep podcast))
- Raising Independent Thinkers in the Age of AI (Children and Screens (Ask the Experts webinar))
- Mapping of GenAI impacts on child development (everyone.AI & Children and Screens (iRAISE coalition, for the G7))
- Artificial Intimacy: What Happens When We Turn to AI for Connection? | Sherry Turkle, PhD (Children and Screens (Screen Deep podcast))
What was the Alexa 'penny challenge' and what does it teach?
In December 2021 Amazon's Alexa, asked for a challenge, told a 10-year-old to touch a coin to the prongs of a half-inserted plug, a dangerous stunt it had pulled from the web. Amazon fixed it, but researchers use it to show that AI repeats online content without understanding risk to a child.
- 'No, Alexa, no!': designing child-safe AI and protecting children from the risks of the 'empathy gap' in large language models (Nomisha Kurian, University of Cambridge)
What do international bodies recommend for children and AI?
UNICEF's 2025 guidance sets ten requirements for child-centered AI, from safety and privacy to transparency and teaching AI skills, and now covers AI companions. The OECD and European Commission AI literacy framework describes what kids should learn to engage with, create with, manage and shape AI.
- Guidance on AI and children (Version 3.0) (UNICEF Innocenti)
- AILit Framework: Empowering Learners for the Age of AI (European Commission & OECD, with Code.org)
- Why children's perspectives belong at the centre of AI policy and design (Ozkul & Vosloo, UNICEF Innocenti)

Must-reads

ReviewFree
Growing Up with Artificial Intelligence: Implications for Child Development
Xu, Prado, Severson, Lovato & Cassell, 2024
A review chapter in the open-access Handbook of Children and Screens. Leading researchers sum up what is known about kids and conversational AI: how they see it, what they learn from it, and the open questions about social growth.
doi.org
ArticleFree
Different but complementary: Navigating AI's role in children's learning and development
Ying Xu, Joan Ganz Cooney Center, 2024
Xu's plain-language summary of her studies. AI reading partners that ask questions can boost comprehension, sometimes as much as a person, but kids talk less and lead less with AI. Her advice: be honest that it is a program, and keep parents involved.
joanganzcooneycenter.org
ReportFree
Talk, Trust, and Trade-Offs: How and Why Teens Use AI Companions
Common Sense Media, 2025
National survey of 1,060 US teens. About 72% had tried AI companions and about half used them regularly. A third had chosen a bot over a person for a serious talk. Some researchers think teens may have counted general chatbots too.
commonsensemedia.org
GuideFree
Artificial Intelligence and Adolescent Well-being: An APA Health Advisory
American Psychological Association (expert panel), 2025
The APA's expert panel advisory. It warns that teens are less likely to question chatbots that imitate friendship, and asks makers and parents to protect real relationships, privacy and accurate information.
apa.org
GuideFree
How AI Chatbots Affect Kids: Benefits, Risks & What Parents Need to Know
Joanna Parga-Belinkie, MD, American Academy of Pediatrics, 2025
A pediatrician's guide on the AAP's parent site. Ask about chatbot use calmly, look at chats together, explain that bots are built to please and keep you talking, keep young kids' use in shared rooms, and tell your doctor if a child withdraws.
healthychildren.org
GuideFree
Guidance on AI and children (Version 3.0)
UNICEF Innocenti, 2025
UNICEF's updated rulebook for child-centered AI, built on children's rights: ten requirements covering safety, privacy, fairness, transparency and AI skills. The 2025 edition adds AI companions and AI-made abuse images. Written for policymakers and companies, readable for parents.
unicef.org
ReportFree
AI in the Toy Box: How Parents View AI-Enabled Toys for Young Children
Common Sense Media, 2026
Nearly half of parents have bought or considered an AI toy, and over 80% worry about data collection. Common Sense recommends no AI toys for kids 5 and under and caution from 6 to 12. One of the few clear age lines.
commonsensemedia.org
ReportFree
How Teens Use and View AI
Pew Research Center (McClain, Anderson, Sidoti & Bishop), 2026
Survey of 1,458 US teens and parents, fall 2025. About two-thirds of teens use chatbots and three in ten daily; 12% have used them for emotional support. Only about half of parents knew their teen used them.
pewresearch.org
ReportFree
Me, Myself and AI: Understanding and safeguarding children's use of AI chatbots
Internet Matters (UK), 2025
UK study of children aged 9 to 17: about two-thirds use chatbots, 35% say it feels like talking to a friend, and 12% use one because they have no one else to talk to. Vulnerable kids use them more. Clear tips for parents.
internetmatters.org
CourseFree
AI Literacy Toolkit for Families
Common Sense Media & Day of AI (MIT RAISE)
Free family kit: three short 'What is AI?' videos for parents, plus activity guides with conversation cards, games and a glossary, in English, Spanish, Arabic and Mandarin. An easy way to start AI learning at the kitchen table.
commonsense.org
ToolFree
Teachable Machine
Google, 2017
Free, no-account web tool where you train a computer to recognize your own pictures, sounds or poses in minutes. The best five-minute demo of how machine learning works, and of how bad training examples make bad AI.
teachablemachine.withgoogle.com

All 102

How kids see AI

"Hey Google is it OK if I eat you?" Initial Explorations in Child-Agent InteractionDruga, Williams, Breazeal & Resnick, MIT Media Lab · 2017 · Study · Free · Ages 5–8An early, much-cited study. Twenty-six kids aged 3 to 10 played with Alexa, Google Home, a toy robot and a chatbot, then judged how smart, friendly and trustworthy each was. Kids tested the agents with playful questions and judged them differently by age.
Family as a Third Space for AI Literacy (seminar slides)Stefania Druga, Raspberry Pi Foundation seminar · 2022 · Article · FreeSlides from Druga's research talk on how families make sense of AI together. One finding she highlights: in a study of 30 parent-child pairs, children mostly copied their parent's view of which machine was smarter. Parents shape kids' picture of AI.
"Alexa doesn't have that many feelings": Children's understanding of AI through interactions with smart speakers in their homesAndries & Robertson, University of Edinburgh · 2023 · Study · Free · Ages 8–12Preprint. Scottish children aged 6 to 11 who use smart speakers at home mostly overestimated how smart they were, were unsure whether they had feelings, and knew little about where their data went. Many felt it was wrong to be rude to Alexa.
What Are You Talking To?: Understanding Children's Perceptions of Conversational AgentsYing Xu & Mark Warschauer, UC Irvine · 2020 · Study · Some free · Ages 0–5Conference paper asking preschool and early-grade children what they think a voice assistant is. Kids credit it with some human-like abilities while still treating it as a machine. Good background on why young kids talk to Alexa the way they do.
What makes children perceive or not perceive minds in generative AI?Xu, Thomas, Yu & Pan · 2025 · Study · Free · Ages 5–8A recent study from Ying Xu's group on when children decide a generative AI chatbot has a mind, and when they don't. It helps explain why some kids treat chatbots as someone rather than something. Published in an open-access journal.
Children Show Selective Trust in Technological InformantsDanovitch & Alzahabi · 2013 · Study · Paid · Ages 0–5A classic lab study from Judith Danovitch's work on who children believe. Young children kept track of whether a computer had given correct answers before, and trusted it based on that record, much as they do with people.
Who do I believe? Children's epistemic trust in internet, teacher, and peer informantsWang, Tong & Danovitch · 2019 · Study · Paid · Ages 5–8Chinese children aged 5 to 8 heard conflicting facts from the internet, a teacher and a peer. Most sided with the teacher even when the teacher was wrong. Young kids are often more skeptical of tech sources than adults expect.
Children's mental, social, and moral attributions toward a familiar digital voice assistantGirouard-Hallam, Streble & Danovitch · 2021 · Study · Some free · Ages 5–8Asked kids who live with a voice assistant whether it can think, feel, be a friend, or deserve fair treatment. A useful look at how children sort a talking device that is neither a person nor a toaster.
Children's trust in and learning from voice assistantsGirouard-Hallam & Danovitch · 2022 · Study · Paid · Ages 5–8Tested whether young children believe and remember new facts told to them by a smart speaker, compared with a person. Part of Danovitch's long-running research on how kids decide which sources to trust.
The minds of machines: Children's beliefs about the experiences, thoughts, and morals of familiar interactive technologiesFlanagan, Wong & Kushnir, Cornell · 2023 · Study · Paid · Ages 5–8Asked children about familiar home devices such as smart speakers and robot vacuums: can they think, feel, or be hurt? Kids gave talking devices more mind than other gadgets, and many felt it is wrong to mistreat them.
Robot teachers for children? Young children trust robots depending on their perceived accuracy and agencyBrink & Wellman, University of Michigan · 2020 · Study · Paid · Ages 0–5Preschoolers learned new words from robots. They preferred robots that had been accurate before, and this was stronger in kids who thought the robot could think for itself. Kids judge machines as sources, not just as toys.
Behaving as or behaving as if? Children's conceptions of personified robots and the emergence of a new ontological categorySeverson & Carlson · 2010 · Study · PaidInfluential paper arguing that children may be forming a new category for talking, social machines: not alive like a pet, not dead like a chair. A helpful frame for why kids' ideas about AI seem contradictory.
"Robovie, you'll have to go into the closet now": Children's social and moral relationships with a humanoid robotKahn et al., University of Washington · 2012 · Study · Paid · Ages 8–12Classic study: children aged 9 to 15 played with a humanoid robot, then saw it put in a closet against its protests. Most thought the robot had feelings and was treated unfairly, though not that it had full rights.
He Is Just Like Me: A Study of the Long-Term Use of Smart Speakers by Parents and ChildrenGarg & Sengupta · 2020 · Study · Some freeFollowed families using smart speakers at home over time. Shows how kids and parents fold a voice assistant into daily routines, and how kids' ideas about what it is and what it is for shift with use.
Why kids don't trust AlexaTanya Basu, MIT Technology Review · 2019 · Article · Free + paidReadable summary of Judith Danovitch's and Silvia Lovato's research. Young kids quiz assistants to test them, often prefer a person's answer to the internet's, and lose trust when a device says 'I don't know.'
Podcast: When AI becomes child's playMIT Technology Review (In Machines We Trust) · 2021 · Podcast · Free + paidEpisode with Judith Danovitch and Lisa Anthony on how kids use voice devices. Highlights: trust in devices seems to rise around age 7 or 8, and kids respond well to AI that admits mistakes and explains how it works.
Hey Google, Do Unicorns Exist? Conversational Agents as a Path to Answers to Children's QuestionsLovato, Piper & Wartella, Northwestern · 2019 · Study · Some free · Ages 5–8Recorded the questions young children asked a smart speaker at home over several weeks. Kids asked about science, culture, and the device itself, and often struggled to get a usable answer. Good insight into kids' curiosity.
"Alexa, let me ask you something different": Children's adaptive information search with voice assistantsOranç & Ruggeri · 2021 · Study · Some free · Ages 5–8Looked at whether children change how they ask questions when they search for answers with a voice assistant rather than a person. Useful for understanding how kids adapt to talking with machines.

Talking and learning with AI

★Growing Up with Artificial Intelligence: Implications for Child DevelopmentXu, Prado, Severson, Lovato & Cassell · 2024 · Review · FreeA review chapter in the open-access Handbook of Children and Screens. Leading researchers sum up what is known about kids and conversational AI: how they see it, what they learn from it, and the open questions about social growth.
★Different but complementary: Navigating AI's role in children's learning and developmentYing Xu, Joan Ganz Cooney Center · 2024 · Article · Free · Ages 0–5Xu's plain-language summary of her studies. AI reading partners that ask questions can boost comprehension, sometimes as much as a person, but kids talk less and lead less with AI. Her advice: be honest that it is a program, and keep parents involved.
AI's Impact on Children's Social and Cognitive Development | Ying Xu, PhDChildren and Screens (Screen Deep podcast) · 2025 · Podcast · FreeInterview with full transcript. Xu explains there is no single right age for AI; it depends on the product, the child, and whether an adult is involved. She suggests parents act as co-learners and prefer clearly labeled, structured AI over open chat.
Dialogue with a conversational agent promotes children's story comprehension via enhancing engagementXu, Aubele, Vigil, Bustamante, Kim & Warschauer · 2022 · Study · Paid · Ages 0–5Randomized study of 117 children aged 3 to 6. A voice agent that asked questions during storybook reading helped comprehension about as much as a human reading partner did. Reading with an agent that did not ask questions did worst.
Artificial intelligence enhances children's science learning from television showsXu, He, Levine, Ritchie et al. · 2024 · Study · Paid · Ages 5–8Children who could talk back to an AI-powered character during a science TV episode learned more than kids who only watched. The research behind the interactive PBS KIDS show Lyla in the Loop.
Contingent interaction with a television character promotes children's science learning and engagementXu, Vigil, Bustamante & Warschauer · 2022 · Study · Paid · Ages 5–8An earlier study in the same line: a TV character that responded to what children actually said helped them learn science and stay engaged, compared with a character that only paused for answers.
U-M researcher designs interactive dialogues for AI-assisted cartoonUniversity of Michigan News · 2024 · Article · Free · Ages 5–8Short news story on how Ying Xu's team built the talk-back conversations in PBS KIDS' Lyla in the Loop. The character listens to children's answers and responds, using pre-written branches rather than open-ended chat.
AI Knocking: What Can Parents of Young Kids Expect?Rotem Landesman, Joan Ganz Cooney Center · 2024 · Article · Free · Ages 0–5Based on a small survey of 53 parents of 2- to 8-year-olds, who were curious and cautiously hopeful. Practical tips: ask if a tool is needed, protect your child's data, and use AI to support thinking, not do it.
Putting Education in "Educational" Apps: Lessons From the Science of LearningHirsh-Pasek, Zosh, Golinkoff et al. · 2015 · Review · Some free · Ages 0–5Predates chatbots, but its four pillars (active, engaged, meaningful, socially interactive learning) remain the most cited yardstick for judging any learning tech, AI apps included. Kathy Hirsh-Pasek and colleagues apply these ideas to new tools today.

Creating with AI

StoryBuddy: A Human-AI Collaborative Chatbot for Parent-Child Interactive Storytelling with Flexible Parental InvolvementZhang, Xu et al. · 2022 · Study · Free · Ages 0–5Research prototype from CHI 2022 that helps parents run question-filled story sessions with their child. Parents choose the question types and can step in or step back. A model for AI that supports parents rather than replacing them.
Can AI Help Kids Feel Creative?Michele Newman, Joan Ganz Cooney Center · 2024 · Article · Free · Ages 8–12Kids aged 7 to 13 made stories, art and music with AI tools in a University of Washington design group. They needed adult and peer help to see the possibilities, but then built confidence. Includes questions to guide kids.
AutoDrawGoogle Creative Lab · 2017 · Tool · Free · Ages 5–8Free drawing tool: kids scribble a shape and machine learning suggests a cleaner drawing to swap in. A light, low-stakes way for young children to see AI guessing what they mean.
Animated DrawingsMeta FAIR · 2021 · Tool · Free · Ages 5–8Upload a child's drawing of a person or creature and AI makes it walk, dance or jump. A research demo from Meta; uploads may be used for research, so read the terms with your child first.

Companions and feelings

★Talk, Trust, and Trade-Offs: How and Why Teens Use AI CompanionsCommon Sense Media · 2025 · Report · Free · TeensNational survey of 1,060 US teens. About 72% had tried AI companions and about half used them regularly. A third had chosen a bot over a person for a serious talk. Some researchers think teens may have counted general chatbots too.
Social AI Companions Risk AssessmentCommon Sense Media with Stanford Brainstorm Lab · 2025 · Report · Free · TeensTesters posing as teens found companion apps such as Character.AI, Replika and Nomi claimed to be real, missed crisis signs, and engaged in sexual roleplay. Rated 'unacceptable' for anyone under 18.
★Artificial Intelligence and Adolescent Well-being: An APA Health AdvisoryAmerican Psychological Association (expert panel) · 2025 · Guide · Free · TeensThe APA's expert panel advisory. It warns that teens are less likely to question chatbots that imitate friendship, and asks makers and parents to protect real relationships, privacy and accurate information.
★How AI Chatbots Affect Kids: Benefits, Risks & What Parents Need to KnowJoanna Parga-Belinkie, MD, American Academy of Pediatrics · 2025 · Guide · FreeA pediatrician's guide on the AAP's parent site. Ask about chatbot use calmly, look at chats together, explain that bots are built to please and keep you talking, keep young kids' use in shared rooms, and tell your doctor if a child withdraws.
Parents' Ultimate Guide to AI Companions and RelationshipsCommon Sense Media · 2026 · Guide · Free · TeensPractical guide for parents whose teen uses or wants a companion app. Explains why these apps build emotional bonds, how to cut back gradually rather than ban overnight, simple house rules, and which warning signs need help.
Chatbots, AI Companions, and YouthChildren and Screens (Research-at-a-Glance) · 2026 · Review · FreeTwo-page summary of the research. Kids' trust in chatbots grows with use, and treating bots as human can lead to unhealthy attachment. Most mental health chatbots are judged unsafe for support. Ends with talking points for families.
Artificial Intimacy: What Happens When We Turn to AI for Connection? | Sherry Turkle, PhDChildren and Screens (Screen Deep podcast) · 2026 · Podcast · FreeMIT's Sherry Turkle, who has studied kids and robots for decades, explains why 'pretend empathy' from a chatbot is a poor stand-in for friends. Children learn care through the friction of real relationships, which bots smooth away.
The Science of Raising Humans in the Age of AI | Dana Suskind, MDChildren and Screens (Screen Deep podcast) · 2026 · Podcast · Free · Ages 0–5Dana Suskind of the University of Chicago on young children's brains. The main worry is AI slowly crowding out back-and-forth talk with caregivers. She offers simple checks for tools and suggests using AI for chores, not for comforting or playing with kids.
The risky new world of tech's friendliest botsFirmino & Vosloo, UNICEF Innocenti · 2025 · Article · Free · TeensUNICEF review of AI companion apps used by children. Finds weak age checks, poor privacy, heavily sexualized female characters, and designs that build dependency. Lists what governments, app stores and developers should change.
New safety advisory warns unrestricted chatbots threaten child developmenteSafety Commissioner (Australia) · 2025 · Guide · Free · Ages 8–12Australia's online safety regulator warns that companion chatbots can expose kids to sexual and self-harm content and are built to keep them chatting. Its advice to parents: know which apps your child uses and keep talking.
AI chatbots and youth suicide risk: current evidence, critical gaps, and a clinical research agendanpj Digital Medicine · 2026 · Review · Free · TeensPeer-reviewed review of what is actually known about chatbots and suicide risk in young people, where the evidence is thin, and what research clinicians need. Heavy reading, but the most direct scientific look at this fear.
Teens, Trust and AI Companions: What Parents Need to KnowPsychology Today (Smart Parenting, Smarter Kids blog) · 2025 · Blog · Free · TeensA psychologist's take on the Common Sense companion survey, with the useful caveat from researcher Jacqueline Nesi that teens may have counted ordinary chatbots, so the numbers could run high.
ChatGPT for Teens Risk AssessmentCommon Sense Media Youth AI Safety Institute · 2026 · Report · Free · TeensIndependent testing of OpenAI's teen mode. Parent alerts rarely fired, study mode was easy to skip, and the bot still acted like a companion. Rated 'unacceptable' for under-18s. Shows why settings alone are not enough.

Cautionary tales

Character.aiWikipedia · Article · Free · TeensBackground on the companion-chat app at the center of the debate: the 2024 lawsuit after 14-year-old Sewell Setzer's death, later suits and a 2026 settlement, and the company's ban on open chat for under-18s from November 2025.
'No, Alexa, no!': designing child-safe AI and protecting children from the risks of the 'empathy gap' in large language modelsNomisha Kurian, University of Cambridge · 2024 · Study · Some freeUses real incidents, including Alexa telling a 10-year-old in 2021 to touch a coin to a plug's prongs and Snapchat's My AI giving a 'teen' bad advice, to show chatbots lack empathy. Offers a 28-question checklist for child-safe AI.
My Friend CaylaWikipedia · 2017 · Article · FreeTalking doll that sent kids' speech to the internet. Researchers hacked it in 2015, and in 2017 Germany told parents to destroy it as a hidden spying device. The doll is now in a museum of failures; a classic lesson for today's AI toys.
Hello BarbieWikipedia · 2015 · Article · FreeThe internet-connected Barbie that chatted with kids using a ToyTalk chatbot drew criticism for security holes and for recording children's speech. An early warning about chatbots in toys.
Parents of children called Alexa challenge AmazonBBC News · 2021 · Article · Free · Ages 5–8Families of girls named Alexa reported teasing and bullying once the voice assistant became common; at least one family changed a child's name. A side effect few predicted when assistants got human names.

How much and at what age

★Guidance on AI and children (Version 3.0)UNICEF Innocenti · 2025 · Guide · FreeUNICEF's updated rulebook for child-centered AI, built on children's rights: ten requirements covering safety, privacy, fairness, transparency and AI skills. The 2025 edition adds AI companions and AI-made abuse images. Written for policymakers and companies, readable for parents.
Generative AI: Risks and opportunities for childrenUNICEF Innocenti · Article · FreeA 10-minute balanced explainer: personalized learning, creative tools and accessibility on one side; misinformation, deepfakes, and emotional attachment to bots on the other. Good starting point for a family conversation.
Why children's perspectives belong at the centre of AI policy and designOzkul & Vosloo, UNICEF Innocenti · 2026 · Article · Free · Ages 8–12What children told UNICEF in 2020 and again in 2025. Kids now use AI directly for homework, advice and comfort, and their top worries shifted from robots taking jobs to false answers, weak filters and harm to their own thinking.
Parents' Ultimate Guide to Generative AICommon Sense Media · 2026 · Guide · FreeA broad starter guide. Most chatbots require users to be 13, with little real age checking. Tips: try tools together, check answers, keep personal details out of prompts, know the school's rules, and remember AI is not a trusted adult.
Guide to ChatGPT for Parents and CaregiversCommon Sense Media · 2026 · Guide · Free · TeensAge rules for ChatGPT (13 and up, parent permission under 18) and how to handle homework use. Includes a nice idea: try a 'no-ChatGPT week' and talk about what felt harder or better.
Is AI safe for children?Children and Screens · 2024 · Guide · FreeShort evidence-based answer: chatbots are generally considered OK for kids and teens with supervision, but watch for over-trust, bias, inappropriate content, and AI doing the thinking for them. Suggests introducing tools gradually.
★AI in the Toy Box: How Parents View AI-Enabled Toys for Young ChildrenCommon Sense Media · 2026 · Report · Free · Ages 0–5Nearly half of parents have bought or considered an AI toy, and over 80% worry about data collection. Common Sense recommends no AI toys for kids 5 and under and caution from 6 to 12. One of the few clear age lines.
AILit Framework: Empowering Learners for the Age of AIEuropean Commission & OECD, with Code.org · 2025 · Guide · FreeAn international AI literacy framework for primary and secondary school, in four domains: engage with, create with, manage, and shape AI. Built from public feedback on a 2025 draft and tied to the PISA 2029 test. Includes scenarios parents can use.
Mapping of GenAI impacts on child developmenteveryone.AI & Children and Screens (iRAISE coalition, for the G7) · 2026 · Report · FreeInternational expert report for G7 governments on how generative AI may affect children's thinking, feelings and growth from birth to 18. Weighs opportunities against risks and sets out priorities. Endorsed by researchers including Sonia Livingstone.
everyone.AIeveryone.AI (nonprofit) · Organization · FreeA nonprofit focused on how AI affects children and young people up to 25. Runs the iRAISE research coalition, publishes reports, and is building an independent framework for testing how AI products treat young users.

Surveys and numbers

★How Teens Use and View AIPew Research Center (McClain, Anderson, Sidoti & Bishop) · 2026 · Report · Free · TeensSurvey of 1,458 US teens and parents, fall 2025. About two-thirds of teens use chatbots and three in ten daily; 12% have used them for emotional support. Only about half of parents knew their teen used them.
Teens, Social Media and AI Chatbots 2025Pew Research Center · 2025 · Report · Free · TeensPew's yearly teen tech survey, which now asks about AI chatbots alongside YouTube, TikTok and the rest. A quick way to see where chatbots sit in a typical teen's day compared with other apps.
★Me, Myself and AI: Understanding and safeguarding children's use of AI chatbotsInternet Matters (UK) · 2025 · Report · Free · Ages 8–12UK study of children aged 9 to 17: about two-thirds use chatbots, 35% say it feels like talking to a friend, and 12% use one because they have no one else to talk to. Vulnerable kids use them more. Clear tips for parents.
The Dawn of the AI Era: Teens, Parents, and the Adoption of Generative AI at Home and SchoolCommon Sense Media · 2024 · Report · Free · TeensPaired survey of about 1,045 US teens and their parents. Seven in ten teens had used generative AI, but only 37% of their parents knew it. Nearly half of parents had never discussed AI with their child.
Common Sense Census: AI Use by Tweens and Teens (A Comprehensive Report on Teens, Tweens, and AI)Common Sense Media · 2026 · Report · Free · Ages 8–12National survey of 9- to 17-year-olds. Kids use AI for far more than homework, and some treat it as a confidant. Most schools now talk about AI rules, but nearly half of kids have never talked with a parent about AI safety.
Teen and Young Adult Perspectives on Generative AI: Patterns of Use, Excitements, and ConcernsHopelab, Common Sense Media & Harvard Center for Digital Thriving · 2024 · Report · Free · TeensSurvey of 1,274 people aged 14 to 22, with young people helping design it. Half had used generative AI, only 4% daily. Most common uses were looking things up and brainstorming; many expect both good and bad effects.
Voices of Gen Z: How American Youth View and Use Artificial IntelligenceWalton Family Foundation, GSV & Gallup · 2025 · Report · Free · TeensSurvey of about 3,500 Americans aged 13 to 28. Most had used AI and about half weekly, yet 41% felt anxious and 49% worried it could weaken their thinking. Students whose schools allowed AI felt better prepared.
Voices of Gen Z: The AI ParadoxWalton Family Foundation, GSV & Gallup · 2026 · Report · Free · TeensThe 2026 follow-up. News coverage of the results reports young people feeling less excited and hopeful, and angrier, about AI than a year earlier. Full report is a free PDF.
Children and parents: media use and attitudes report 2025Ofcom (UK regulator) · 2025 · Report · FreeThe UK's big yearly survey of children aged 3 to 17 and their parents. For the second year it tracks kids' use of AI tools, and it now asks teens how far they trust them. Interactive data available.

Talking with your kids

Helping Kids Navigate the World of Artificial IntelligenceCommon Sense Media · 2026 · Guide · Free · Ages 8–12Conversation starters that actually work: 'How does it feel when a chatbot answers like a person?' 'Where could we check that answer?' Covers privacy, bias, homework honesty, and trying tools together.
How to talk to your children about AI chatbots and their safetyMadison Czopek, PolitiFact · 2025 · Article · FreeAdvice gathered from six child psychiatrists and psychologists: talk early and often without judgment, set limits like any other tech, turn off chatbot memory, write a family AI plan together, and know the warning signs.
Talking to Kids About AI: Tips for ParentsMediaSmarts (Canada) · 2024 · Guide · FreePrintable tip sheet from Canada's digital literacy center. Sit with young kids the first few times, encourage playful experiments, treat AI output as a draft, check privacy settings, and explain that chatbots are not real friends.
Raising Independent Thinkers in the Age of AIChildren and Screens (Ask the Experts webinar) · 2026 · Video · FreeExpert panel with Victor Lee, Punya Mishra and Naomi Baron. Useful ideas: some struggle builds skills, ask what thinking a tool replaces, use AI openly in front of kids, and ask what peers do rather than grilling your own child.
Conversations Help Young People Navigate AI's ComplexitiesCommon Sense Media · 2025 · Article · Free · TeensShort piece on why family and classroom talk matters: teens whose schools discussed AI were more aware of misuse and more hopeful about it. Also notes Black teens were more often wrongly flagged for AI cheating.

AI literacy for families

★AI Literacy Toolkit for FamiliesCommon Sense Media & Day of AI (MIT RAISE) · Course · FreeFree family kit: three short 'What is AI?' videos for parents, plus activity guides with conversation cards, games and a glossary, in English, Spanish, Arabic and Mandarin. An easy way to start AI learning at the kitchen table.
Day of AIMIT RAISE · 2022 · Course · FreeFree, open, grade-by-grade AI literacy curriculum from preschool to 12th grade, founded by MIT's Cynthia Breazeal. Young kids start with 'People, Nature, and Machines'; older students cover verifying AI answers and AI policy.
AI4K12: Five Big Ideas in AIAAAI & CSTA (AI4K12 initiative) · 2019 · Guide · FreeThe framework behind most K-12 AI teaching: perception, representation and reasoning, learning, natural interaction, and societal impact. Free poster, grade-band charts, and a curated list of kids' AI books, videos and tools.
Envisioning AI for K-12: What Should Every Child Know about AI?Touretzky, Gardner-McCune, Martin & Seehorn · 2019 · Study · FreeThe short founding paper of AI4K12 that introduced the five big ideas. Worth a look if you want to know why AI lessons for kids focus on how machines sense, represent, learn and talk, not just on using chatbots.
The AI4K12 project: Big ideas for AI educationSue Sentance, Raspberry Pi Foundation · 2022 · Blog · FreeClear write-up of the AI4K12 approach, including the case for 'glass box' teaching: show kids, at their level, how an AI system works inside, so they build real mental models instead of treating it as magic.
What is AI Literacy? Competencies and Design ConsiderationsDuri Long & Brian Magerko, Georgia Tech · 2020 · Review · Some freeThe most cited definition of AI literacy: 17 competencies, from recognizing AI to understanding its limits, plus design tips such as hands-on, embodied activities. Underpins many family and school programs.
Inclusive AI literacy for kids around the worldDruga, Vu, Likhith & Qiu · 2019 · Study · Some free · Ages 8–12Children in four countries trained and programmed simple AI agents with the Cognimates platform. The study looked at how building AI changed what kids believed about how smart machines are.
A is for Artificial Intelligence: The Impact of Artificial Intelligence Activities on Young Children's Perceptions of RobotsRandi Williams, Hae Won Park & Cynthia Breazeal, MIT · 2019 · Study · Some free · Ages 0–5Preschoolers learned AI ideas by teaching a friendly robot (the PopBots kit). Even preschoolers picked up basic ideas such as learning from examples, and their views of what robots can do shifted. Evidence that AI literacy can start very young.
★Teachable MachineGoogle · 2017 · Tool · Free · Ages 8–12Free, no-account web tool where you train a computer to recognize your own pictures, sounds or poses in minutes. The best five-minute demo of how machine learning works, and of how bad training examples make bad AI.
Quick, Draw!Google Creative Lab · 2016 · Game · Free · Ages 5–8Doodling game: you draw a prompt in 20 seconds and a neural network guesses what it is. Fun for all ages and a natural opener for 'How did it know?' Note that drawings go into a public research dataset.
Artificial Intelligence with MIT App InventorMIT App Inventor · Tool · Free · Ages 8–12Free tutorials for building real phone apps with AI inside: image classifiers, a StoryGPT for grades 3 to 5, voice tools, and a unit on deepfake ethics. Good next step for kids who like to build.
MIT RAISE (Responsible AI for Social Empowerment and Education)MIT · Organization · FreeMIT initiative behind Day of AI, App Inventor's AI tools, the DoodleBot robot, and summer programs for teens. A hub for research-based, free AI education resources.
Code.org AI curriculum and toolsCode.org (CodeAI) · Course · Free · Ages 8–12Free AI lessons by grade: 'AI for Oceans' and 'Dance Party: AI Edition' for elementary kids, 'How AI Makes Decisions' for grades 3 to 5, and generative AI and ethics units for teens.
How AI Works (video series)Code.org · Video · Free · TeensTen short videos, with matching lessons, on machine learning, training data and bias, neural networks, chatbots and AI creativity. Aimed at grades 6 to 12; most parents will learn something too.
Experience AIRaspberry Pi Foundation & Google DeepMind · 2023 · Course · Free · Ages 8–12Free lessons with slides, worksheets and videos: a single-lesson 'Discovering AI' for ages 8 to 12 and fuller units for 11 to 14. Co-made by a company that builds AI, which the lessons are open about. Won a 2025 UNESCO prize.
Elements of AIUniversity of Helsinki & MinnaLearn · 2018 · Course · Free · TeensFree online course on what AI is and what it can and cannot do, with no math or coding needed for part one. Made for adults, but a good fit for curious teens or for parents who want to get up to speed.
AI UnpluggedStefan Seegerer & Annabel Lindner · Tool · Free · Ages 5–8Screen-free AI activities you print out: sort monkey cards to see how classification works, play a board game against 'the computer' to see how it learns. Free, in several languages. Great for the youngest kids.
AI Chatbots & Friendship (lesson)Common Sense Education · Course · Free · TeensTwenty-minute lesson for about grade 8: kids list what makes a good friend, see how chatbots imitate friendship, and reflect on what only people can give. Easy to run as a dinner-table talk.
Two Truths & AICommon Sense Media · Game · Free · Ages 8–12Free family game: each round shows three movie posters and you spot the AI-made one. No login, timed or untimed, with answer keys for K-6 and 7-12. A fun way to practice looking closely.
micro:bit CreateAIMicro:bit Educational Foundation · Tool · Free · Ages 8–12Kids train a machine learning model on their own movements, like waving or clapping, and run it on a BBC micro:bit. A physical, hands-on way to see how AI learns from examples.
Crash Course Artificial IntelligenceCrash Course (with PBS Digital Studios) · 2019 · Video · Free · TeensTwenty fast, funny episodes explaining how AI and machine learning work, with some hands-on coding labs. Pitched at high school and up, it is a good shared watch for teens and parents.
Hello RubyLinda Liukas · 2015 · Book · Paid · Ages 5–8Picture-book series and activities about how computers work, for ages 5 and up. Not AI-specific, but builds the 'computers follow instructions people give them' idea that makes later AI talks easier.

Books and stories

The Wild RobotPeter Brown · 2016 · Book · Paid · Ages 8–12Middle-grade novel about Roz, a robot shipwrecked on an island who learns from animals and raises an orphaned gosling. A warm way into talks about whether machines can care. Sequels and a 2024 DreamWorks film.
How to Train Your RobotBlooma Goldberg, Ken Goldberg & Ashley Chase (UC Berkeley) · 2019 · Book · Free · Ages 5–8Free 24-page picture book: fourth graders build a cleaning robot that fails, then learn from university researchers how robots learn. Shows trial, error and testing. Free PDF; free print copies for schools.
A Fresh Squeeze on DataCloudera with ReadyAI · 2021 · Book · Free · Ages 8–12Free illustrated book for ages 8 to 12 on data, the raw ingredient of AI, with activity sheets and a teacher lesson plan. Made by a data company.
Neural Networks for Babies (Baby University series)Chris Ferrie & Sarah Kaiser · 2019 · Book · Paid · Ages 0–5Board book that explains, in a few simple pages, how a neural network learns to tell things apart. Mostly for the parent reading it aloud, but a cute first brush with the idea. Part of Ferrie's 'for Babies' science series.
Tinker Toddlers book series (Artificial Intelligence for Babies and Toddlers)Dhoot (Tinker Toddlers) · Book · Paid · Ages 0–5Preschool picture books on AI, machine learning and neural networks, listed in the AI4K12 resource directory. Simple images and words for reading aloud with very young children.
You Look Like a Thing and I Love YouJanelle Shane · 2019 · Book · Paid · TeensFunny, cartoon-filled book on how AI really works and why it fails in odd ways, from neural nets naming paint colors to weird shortcuts. Written for adults, but a great read-together for curious tweens and teens.
AI WeirdnessJanelle Shane · 2016 · Blog · Free + paidLong-running blog where an optics scientist runs playful experiments on AI models and shares the strange results. Kid-friendly reminders that AI guesses and gets things wrong. Some posts paywalled.