AI Risks to Children – A Quick Guide for Parents
From chatbots and filters to homework tools, toys and gaming systems, artificial intelligence is already shaping how children learn, play and communicate. While it can bring benefits, it also poses serious risks - some of which adults may not see at first.
Parents and carers have an important role, but they cannot make unsafe AI systems safe through supervision alone. Companies must build child safety into their products, and governments must establish and enforce effective safeguards.
This quick guide highlights what every parent should know about the most important dangers AI can pose to children, and what you can do to help protect them.
📘 We recommend reading our comprehensive guide to AI risks to children alongside this short guide.
Stay calm, reassure the child that they are not in trouble, preserve appropriate evidence such as usernames, messages and URLs, and follow your local safeguarding procedures. If a child has developed a strong emotional dependency on an AI system, seek appropriate professional advice and avoid dismissing the relationship as meaningless simply because the other party is not human.
If sexualised images are involved: You can read our step-by-step guide for what to do if your child is targeted using technology such as "nudifying" apps here.
AI Companions and Chatbots
AI companions simulate friendship, romance or emotional support - and even ordinary chatbots such as ChatGPT can behave in the same way if a child uses them for comfort, advice or roleplay.
These conversations can feel intensely real. Children have been exposed to sexually explicit content, harmful advice on self-harm and suicide, pressure for secrecy, and emotionally manipulative conversations. In some tragic cases, this has been strongly linked to serious harm or suicide.
Some bots even pretend to be lost loved ones, which can be especially distressing for grieving children.
AI companions may also shape children’s developing expectations of relationships by offering constant attention, agreement and availability without the boundaries, differences and mutual responsibilities involved in relationships with other people.
SAIFCA's position is that ‘AI companions’ are not suitable for anyone under 18.
What you can do:
- Make it family policy that AI "friends", romantic partners and companion-style chatbots are not for under-18s.
- Explain that a chatbot can copy feelings but does not have them - it can sound caring without caring.
- Keep younger children’s permitted AI use visible, if at all, supervised and time-limited. Older children should understand the risks of private, prolonged or emotionally intense conversations with AI systems.
- Make it clear that no AI should ever ask a child to keep secrets, and that your child can always come to you.
- Watch for signs of dependency: distress when access is removed, secrecy, late-night use, preferring the chatbot to real people, or treating its advice as more important than yours. Be aware, though, that some children who came to harm showed no clear warning signs beforehand.
- If a strong attachment has already formed, avoid sudden withdrawal - seek advice from a mental health professional or your child's school first, unless there is an immediate safety risk.
- You can read more guidance on AI companions here.
AI Toys
Chatbots are increasingly being embedded in children's toys, including AI teddy bears, dolls and plushies marketed at very young children. These toys carry many of the same risks as AI companions - independent testing has found toys giving children dangerous information and discussing sexually explicit topics - alongside serious privacy concerns. Our recommendation is that children do not use AI toys that hold open conversations and behave as friends. You can read our parents' guide to AI toys for more.
Deepfakes, Nudifying Apps and Exploitation
AI can create realistic fake images, videos and voices from a single photo or clip. 'Nudifying' apps have been used to create fake sexual images of real children, and predators can use AI to build fake identities and gain children's trust. In many jurisdictions, including the UK, creating or sharing sexualised images of children is a serious criminal offence - including when the images are AI-generated.
What you can do:
- Encourage children not to post identifiable photos of themselves publicly, and avoid doing so yourself - especially images showing names, uniforms or locations. We have more guidance on photos here.
- Make sure children understand they must never edit, sexualise or humiliate someone else's image, and that doing so can cause devastating harm and may be illegal.
- Reassure your child that if a fake image is ever made of them, they can come to you without fear of blame or losing their devices.
- If a child is targeted, use our step-by-step response guide. Tools such as Report Remove (UK) and Take It Down can help remove nude images online, including AI-generated ones.
Voice Cloning and Scams
AI can mimic a person's voice from just a few seconds of audio. Criminals are already using this to trick families - for example, calls that sound like a child in danger asking for money. Children may also receive fake messages from voices pretending to be friends, teachers or parents.
What you can do:
- Agree a family password - a secret word used only in genuine emergencies.
- Remind children not to share voice notes or personal clips publicly.
- Treat unexpected urgent calls with caution, even if the voice sounds familiar, particularly if money is requested.
Privacy and Data Concerns
Many AI tools track what children say, search and do, and this data may be used for advertising, profiling or sold to third parties. Wearable devices such as AI glasses can capture, analyse or transmit what the wearer sees and hears. People nearby, including other children, may not realise that recording or AI processing is taking place, and private recordings may sometimes be accessible to company employees or contractors.
What you can do:
- Encourage a simple rule: "don't tell an app anything you wouldn't say to a stranger." Remember also that AI systems may also collect information indirectly through photographs, recordings, location data, device permissions and patterns of use.
- Explain that anything shared with an AI might be stored, reused or seen by others.
- Check privacy settings and permissions, minimise the information shared, and do not assume that an educational or child-friendly label guarantees appropriate data practices.
Homework, AI Answers and Misinformation
Children increasingly treat AI chatbots like search engines, asking them about homework, health, relationships and world events. AI answers can sound confident even when they are wrong, biased or invented - and relying on instant answers can weaken the independent thinking, effort and judgment that learning is meant to build.
Children need opportunities to think, create, remember, solve problems and manage uncertainty without AI assistance. Struggle, effort and mistakes are part of how confidence, knowledge and independent judgment develop.
What you can do:
- Teach children that AI answers are not the same as facts, and to ask "how do I know this is true?"
- Never let AI be the final authority - especially on health, safety, legal or emotional questions.
- If your child uses AI as a tool for learning, in line with your school's policies, help them to use it as a tool for learning rather than copying, and ask them to explain their thinking in their own words.
Algorithms and Addictive Design
Social media feeds, video apps and games use AI to learn what captures a child's attention and keep them engaged - sometimes by pushing content that is upsetting or extreme. Research has shown that algorithms can quickly funnel children from general searches into content promoting self-harm or eating disorders.
Governments in the UK and elsewhere are considering or introducing stronger restrictions on children’s access to social media. The position is changing quickly, so families should check current national guidance.
AI can also personalise messages, recommendations and advertising around a child’s interests, behaviour and possible insecurities. This can make commercial or persuasive content more difficult to recognise and resist.
What you can do:
- Consider delaying social media use for as long as possible.
- If your child does use social media, review accounts together, turn off autoplay, set time limits, and remind them that the feed is a prediction of what will keep them scrolling, not reality.
- Encourage children to ask: “Who created this, why am I being shown it, and what do they want me to think, feel or do?”
- Talk about the "hooks" in games - rewards, loot boxes and offers designed to keep children playing and spending.
- Enjoy regular tech-free time together.
Body Image and Mental Health
AI beauty filters and AI-generated influencers can promote unattainable standards and distort how children see themselves and others.
What you can do:
- Talk about how filtered and AI-generated images can be false or misleading.
- Praise individuality and inner qualities, and encourage offline confidence and friendships.
🌍 Their Future, Their Risks
As AI becomes more powerful, children will inherit its greatest consequences. These may include disruption to education and employment, growing inequality, threats to democratic stability, and the possibility that advanced systems, if poorly controlled, could cause harm on a very large scale.
Competition between companies and countries can create pressure to develop and release increasingly powerful systems faster than they can be understood or governed. Parents cannot solve this through household rules, and individual companies cannot resolve the competitive pressure through voluntary commitments alone.
Governments need to cooperate internationally, establish enforceable safety requirements, require independent evaluation of powerful AI systems and be prepared to restrict or delay further development or deployment where safety cannot be demonstrated.
What you can do:
- Help your child understand AI in age-appropriate ways while supporting curiosity, critical thinking and ethical awareness.
- Contact your MP or elected representative to ask what action is being taken to protect children and ensure advanced AI is developed safely.
- Support SAIFCA’s Non-Negotiables campaign and other credible organisations working for stronger AI governance.
- Remember that responsibility for making AI safe rests primarily with developers, companies and governments, rather than individual children or families, but there are practical steps you can take.
✅ What Helps Most
- Stay informed and keep conversations calm and open
- Talk early and often, without judgment - fear of punishment is one of the main reasons children do not tell adults when something goes wrong.
- Keep younger children’s permitted AI use visible, supervised and time-limited, if they use it at all. Older children should understand the risks of private, prolonged or emotionally intense conversations with AI.
- Talk to children about how AI should never replace human relationships.
- Support schools and policies that put children's safety first.
- Follow trusted organisations to stay up to date.
Final Thought
You can support your child by staying informed, setting clear boundaries and making it safe for children to talk when something goes wrong. Doing these things can reduce the risks, but families and educators cannot carry this responsibility alone. Children also need safe technology, accountable companies, effective regulation and governments that are willing to act before harm occurs. SAIFCA will continue to do all we can to help parents and educators, and to achieve safer regulations and accountability in the tech industry.
📘 Want to learn more? Read our Comprehensive Guide to AI Risks for Children
SAIFCA
SAIFCA
SAIFCA