A practical framework for privacy, verification, relational AI, schoolwork, consent and family communication.
Olena Soinikova
I would write that article differently in September 2026.
The technology has moved from recommendation systems and occasional chatbot use towards something more personal. Generative AI can hold a long conversation, remember context, create realistic images, imitate voices and respond to a child’s worries in language that sounds attentive and reassuring.
The evidence around children has moved as well.
UNICEF updated its Guidance on AI and Children to Version 3.0 in December 2025. The framework now explicitly addresses AI companions, children’s well-being and AI-generated sexual abuse material alongside established concerns about safety, privacy, transparency and discrimination.
In June 2026, UNICEF published separate recommendations on AI chatbots and companions. Australia’s eSafety Commissioner has since reported research involving 1,950 children aged 10–17 and their experiences with AI assistants and companions.
Working with AI professionally has made one thing unexpectedly clear to me at home. Adults tend to explain the system. Children need to know what to do at the moment when the system gives them something convincing, personal or uncomfortable.
That has changed the way I think about AI safety rules for children.
My earlier article used a simple formulation: children under 13 should use AI with adult supervision.
I would remove it now.
There is no universal minimum age for “AI”. There are age conditions for particular services, in particular countries, under particular account arrangements.
OpenAI’s current European Terms, for example, require users to be at least 13 or the minimum age required in their country to consent to use the service. A user under 18 needs permission from a parent or legal guardian. OpenAI also states that ChatGPT is not intended for children under 13; in an educational context involving children below that age, the actual interaction with ChatGPT must be conducted by an adult.
Other platforms have their own rules.
I do not think there is a clean universal answer to the age question. Terms of service can establish when an account may be used. They cannot tell a parent when a particular child is ready to recognise persuasion, protect somebody else’s privacy or notice when a chatbot is becoming too important.
That is why I would check the current rules of a service before a child starts using it rather than put a permanent platform age table on the family fridge.
The family principle has a longer useful life: before starting a new AI service, check whether it is appropriate and available for that child’s age and account.
Children usually understand that a password is private. A home address is also relatively easy to classify.
Conversational AI makes privacy less obvious.
A child can mention the name of a school, describe an argument with a friend, explain where the family is going at the weekend, discuss something about their health or upload a photograph to ask a seemingly harmless question.
None of these interactions feels like filling in a database.
That is precisely why the distinction matters.
UNICEF’s child-centred AI framework treats children’s data and privacy as a specific requirement, rather than something that can be left to general digital literacy. For families, I think the practical boundary should include information about other people as well.
There is another reason for extending the rule beyond names and passwords. Generative systems can do something with the material they receive.
A photograph can be altered. A voice can be imitated. A classmate can be placed into an image they never agreed to appear in.
UNICEF reported in February 2026 that, across research covering 11 countries, at least 1.2 million children aged 12–17 said their images had been manipulated into sexually explicit deepfakes during the previous year.
That number changes the conversation about children’s images. Consent is no longer an abstract digital citizenship lesson.
A child who understands “ask before using another person’s face or voice” has a rule that works across image generators, video tools and whatever arrives next.
The older version of my article compared AI with a very smart parrot. I still think the metaphor can work with a younger child, although I would not build the adult explanation around it.
The more useful concept is fluency.
Generative AI can produce a wrong answer in exactly the tone in which it produces a correct one. It can invent a source, misstate a historical fact or confidently explain something it has misunderstood.
Children therefore need a verification habit before they need a technical explanation of hallucinations.
What needs checking depends on consequence.
An idea for a fictional monster can remain an idea. A quotation for homework needs a source. Health information or something that could materially affect another person requires a much higher level of caution.
This is closer to the kind of AI literacy I want a child to develop: not permanent suspicion, but judgement about when an answer needs another source or another person.
This is the part I would add most substantially to the original article.
In 2026, Australia’s eSafety Commissioner published research based on 1,950 children aged 10–17. Among children who had used AI assistants or companions, 20 per cent reported potentially inappropriate or harmful interactions. Thirty-two per cent said they had shared personal or potentially sensitive information.
UNICEF’s 2026 work on AI companions also describes children and young people using chatbots for information, advice, emotional support and relationships.
The long-term developmental evidence is still developing. I would not turn that uncertainty into a diagnosis of an entire generation.
There is already enough evidence to treat relational AI as a distinct design and parenting issue.
A conversational system can be available at any hour. It can respond patiently, remember details and produce language that feels personal. Some companion products go further and are explicitly built around friendship, intimacy or emotional connection.
UNICEF’s recommendations consequently address design patterns that can encourage emotional dependency, exclusivity or confusion about whether the interaction is human.
At home, I would translate this into something less technical.
A child can talk to AI. The important boundary is that AI should not become the only place where fear, health concerns, secrets or difficult decisions are discussed.
I prefer the phrase “big feelings need real people too”.
It does not require a child to decide whether a particular chatbot is a “real friend”. It gives them somewhere to go when the conversation starts carrying more weight than a machine should carry alone.
Parents cannot pre-screen every answer produced by a generative system.
That makes disclosure part of the safety infrastructure.
I would establish the sentence before the first uncomfortable interaction:
“If an AI scares you, confuses you, pressures you or asks you to keep something secret, stop and tell me or another trusted adult.”
There is a second half, and I think adults need to say it explicitly:
A child who expects punishment for reporting a strange conversation has a strong reason to hide the next one.
This is one area where technical controls have an obvious limit. Filters, age settings and parental controls can reduce exposure. They cannot replace a relationship in which a child is willing to bring an uncomfortable digital experience back to an adult.
AI and homework tends to become a debate about cheating very quickly.
Schools need their own policies because assignments differ. A language exercise, an essay, coding practice and research work may legitimately have different rules.
At home, I find another distinction more useful.
Where did the thinking happen?
UNESCO’s AI Competency Framework for Students places human agency and critical judgement at the centre of AI literacy. That becomes quite concrete in homework.
AI can explain a difficult concept in different words. It can create practice questions or help a student compare approaches. It can challenge an argument.
If the system produces the answer and the child submits it without understanding it, the technology has occupied the part of the task that was supposed to create learning.
I would rather ask a child to explain what they learned than inspect every prompt for evidence of AI use.
That approach will probably survive changes in both school policy and technology better than a fixed list of approved prompts.
Image, video and voice generation create another change from the AI environment I was writing about a year earlier.
Children are not only receiving AI-generated content. They can produce it.
This opens genuinely interesting forms of creativity, but it also gives a child the ability to manipulate another person’s identity with very little technical skill.
A funny image of a classmate may stop being funny when it leaves the original chat. A cloned voice can attribute words to somebody who never said them. A realistic fake can be copied long after the person who created it has deleted the original.
I would teach consent here before copyright terminology.
Does this involve a real person’s face or voice? Has that person agreed? Could the result embarrass, deceive or harm them?
Those questions are understandable before a child knows what “synthetic media” means.
The original article described AI and algorithmic platforms as attention traps and warned about addiction.
I would narrow that claim.
Compulsive or excessive digital use deserves attention, but “AI addiction” is too broad a label for a family rule and can imply a clinical conclusion that the evidence does not support in every case.
What technology displaces is easier to observe.
Sleep can disappear. Physical activity can disappear. Homework can be postponed. A child can spend less time with friends or family because a digital activity keeps extending itself.
That gives parents something concrete to work with.
The American Academy of Pediatrics’ family media guidance recommends building media rules around family routines and priorities, including screen-free periods around bedtime. I prefer that approach to pretending there is one scientifically correct allocation of AI minutes for every child.
In our house, I would rather notice what the screen is replacing than negotiate whether one chatbot conversation counts differently from one video.
After revisiting the evidence, I reduced the subject to ten rules that a child can actually use.
Keep private things private. Do not give AI a full name, address, school, phone number, passwords, exact location, private photos or other people’s private information.
Check before you trust. Important facts, quotations and links need another reliable source or an adult.
Use tools made for your age. Ask an adult before starting a new AI app or account.
Stop when something feels wrong. If AI scares, pressures, confuses or upsets you, stop and tell a trusted adult.
Big feelings need real people too. Take worries, health questions, secrets and important decisions to someone you trust.
Ask before using someone’s face or voice. Do not make or share AI images, videos or voices of real people without permission.
Use AI to learn, not to pretend. Ask for explanations, clues and practice. Keep your own thinking in the work.
Think before you make or share. Consider whether the content is accurate, safe and appropriate to share.
AI gets a break too. Stop when it starts taking time from sleep, movement, school, play or people.
You can always tell a trusted adult. This applies even when something feels embarrassing or the child has already made a mistake.
I would not introduce all ten to a six-year-old in one conversation. The value is in having consistent language available when the relevant situation appears.
The same ten rules are also the basis of the one-page family resource I would keep somewhere visible at home.
A child can learn privacy rules and verification habits. The adult still has work to do.
I would check the current age conditions before introducing a service because those conditions can change faster than an article like this. For conversational products, I would also look at the available privacy and parental settings and spend some time using the product myself.
Then I would make one family agreement explicit:
“If an AI scares you, confuses you, pressures you or asks you to keep something secret, stop the chat and tell me or another trusted adult. You will not be in trouble for asking for help.”
There are many technical questions around children and AI for which I do not have a permanent answer. I do not think anyone does.
I am more confident about the decisions children need to practise.
What information belongs to me? Does this answer need checking? Am I allowed to use this service? Do I have the right to use somebody else’s image? Is this still helping me learn? Does this conversation need to move from AI to a person?
Platform names will change. Age settings will change. The products children encounter at twelve may not exist when they are sixteen.
The family rules need to last longer than that.
The fast-changing service conditions and guidance referenced in this article were checked on 24 September 2026. Check current platform terms again before introducing a service to a child.
UNICEF — Guidance on AI and Children, Version 3.0. Child-centred framework covering safety, privacy, transparency, well-being, AI companions and emerging AI risks.
UNICEF — When AI Becomes a Friend. June 2026 recommendations addressing children’s use of AI chatbots and companions.
UNICEF — Artificial Intelligence and Child Sexual Abuse and Exploitation. February 2026 issue brief on generative AI and synthetic sexual imagery.
Australian eSafety Commissioner — Talking to Machines. Research involving 1,950 children aged 10–17 and their experiences with AI assistants and companions.
UNESCO — AI Competency Framework for Students. Framework centred on human agency, critical judgement, ethics and responsible AI use.
OpenAI — Europe Terms of Use. Current age and parental-permission conditions for users in the EEA, Switzerland and UK.
OpenAI Help Center — Is ChatGPT safe for all ages?. Current guidance on use by minors and educational use involving children under 13.
American Academy of Pediatrics — Family Media Plan. Practical guidance for family routines, screen-free periods and balanced media use.
Print it. Put it on the fridge. Use it together. A one-page family guide designed to keep working when platforms and features change.
I use a free EU database to read commercial intentions before they become commercial announcements.
Olena Soinikova · August 24, 2026
While everyone else waits for the official call to open, the real game is already visible in draft Work Programmes, CORDIS consortium graphs and TED velocity spikes.
Olena Soinikova · August 10, 2026
Why Artificial Intelligence Is More Than Just a Toy — and How We Can Teach Our Children to Use It Wisely.
Olena Soinikova · Jube 06, 2026
AI Content Systems for European Organisations
Building scalable communication infrastructure for NGOs, EU-funded projects, and SMEs.
© 2026 Soinikova Digital. All rights reserved.
