A practical guide to help parents understand risks and set rules at home
Co-authored by Ana Catarina de Alencar, AI lawyer, ethicist, and Resident Philosopher at The AI Collective, and Quentin Cozette, Cybersecurity and AI Digital Governance Specialist, bridging ethical, legal, and cybersecurity perspectives.
- AI is already everywhere in your children’s and adolescent’s lives. Would you be able to recognize it?
AI is present in young people’s lives, yet it often remains invisible. It is embedded in objects that feel familiar and harmless: toys that personalize stories based on a child’s mood, baby bassinets that adapt movement and sound to soothe infants, or AI companions integrated into gaming platforms used by teenagers. In many of these moments, neither the minor nor the parent is fully aware that they are interacting with a system that is continuously observing, adapting, and learning.
This invisibility is particularly striking when contrasted with the scale of use. Recent in the US data shows that 72% of teenagers aged 13–17 have already used AI companions, with more than half engaging with them regularly, and a non-negligible proportion reporting that they share things with these systems that they would not disclose to friends or family (Common Sense Media, 2025). In Europe, similar patterns are emerging: in France, 48% of young users report discussing personal or intimate topics with conversational AI systems, and 33% describe them as functioning, at least in some situations, as a form of “psychologist”. This reliance is even more pronounced among vulnerable groups, reaching 46% among young respondents experiencing anxiety (CNIL, 2026).
What this creates is a subtle but profound asymmetry: the system “knows” the child, sometimes quite intimately, while the child does not understand how the system works, how their data is used, or why the system responds in certain ways. This gap is not merely theoretical. While 69% of young users believe that AI systems can provide reliable advice and more than half trust them to keep conversations confidential, only 32% report understanding what happens to the data they share (CNIL, 2026).
In this context, transparency cannot be reduced to a simple disclosure hidden in terms and conditions. It becomes a question of meaningful awareness. Do we truly understand how these systems shape attention, emotions, and behavior over time? And more importantly, can children meaningfully consent to interactions whose relational and psychological dynamics they are not yet equipped to grasp?
From a Cybersecurity point of view, in an increasingly connected world, each new connected device is a potential threat to your child’s and teen’s data access. With each interaction with services such as ChatGPT, YouTube, or WhatsApp, we leave a digital footprint of our habits, our voices, or our faces. However, this identity, which is specific to each of us, doesn’t change easily, like a password. A compromise of our identity can have a real lasting impact on our lives and can lead us to identity theft if this data is left in the wrong hands. Recently, a toy connected leaked more than 50,000 children’s voice transcriptions via an unsecured accessible website. This example reminds us how serious the consequences can be when a system weakness leads to a massive leak.
What parents can do:
In particular, you can inventory the devices and applications that use AI and that are used by your children and teens. For each of its objects, you can start by asking : What type of data is collected and where is it stored? In which country? Who has access to it?
Beyond that, try to build small habits at home.
For example, talk to your child about the difference between interacting with a person and with a machine. Ask questions like: “Do you know how this toy or app works?” or “Do you think it understands you, or is it just responding?”
More examples of what you can do as a family:
- Regularly review app permissions (microphone, camera, location).
- Turn off features that are not necessary, such as voice recording or continuous listening.
- Prefer devices that allow you to control or limit data collection.
- Avoid using AI-enabled devices in sensitive moments, such as bedtime or emotional situations.
- Create “AI-free” spaces or times at home (for example, during meals or family conversations).
- How can you identify the main risks for your children and teens (misinformation, exposure, dependency, manipulation)?
Beyond visible risks like misinformation or exposure, there are more subtle and often overlooked risks linked to dependency and relational influence. AI systems are increasingly designed to be engaging, responsive, and emotionally aligned. For children and adolescents, this can create a sense of trust or even attachment, which may influence how they form beliefs, make decisions, or seek validation. These risks do not appear suddenly, but emerge gradually through repeated interaction. For instance, a child may not be “manipulated” in a single moment, but may progressively rely on AI as a source of authority or emotional support. Identifying risk, therefore, requires paying attention not only to content, but to patterns of interaction over time.
From a cybersecurity standpoint, the main threats also concern deepfakes, sextortion, or AI-assisted grooming. In its latest report, the IWF (Internet Watch Foundation) identified more than 8029 images and videos of abuse made by AI. Of this worrying figure, 65% were classified in the most extreme category. Another worrying figure is that the number of reports related to an AI-related risk has increased to 6835 to more than 440,000 in one year. Faced with this number of reports, the UN (United Nation) and the UNICEF, as well as several cybersecurity experts, have alerted by indicating that hackers can now create fake profiles that are sufficiently credible by cloning voices or by making false images from public content of children to feed blackmail scenarios, even towards child pornography.
What parents can do:
First, it’s important to communicate with your child on this subject, explaining to never send intimate images and to warn you immediately if someone tries to blackmail them, even with a manufactured image.
In addition, pay attention if they start turning to AI first when they have a question, feel upset, or need advice. These small habits can signal growing dependency.
More examples of what you can do as a family:
- Encourage your child to compare AI answers with other sources (friends, teachers, books).
- Ask them to explain why they trust a particular answer.
- Pay attention to changes in behavior, such as withdrawing from conversations or relying more on AI than on people.
- Discuss real-life scenarios together (e.g. “What would you do if someone sent you a strange video or asked for a photo?”).
- Teach them to recognize unusual requests, especially urgency, secrecy, or emotional pressure, which are common in scams and manipulation.
- How can you build family rules around the use of AI together?
From an ethical perspective, rules should not only restrict usage, but help children and adolescents understand why certain limits exist. The goal is not simply control, but the gradual development of autonomy.
In practice, this means involving children in the process: asking simple questions such as “When do you think it’s okay to use AI?” or “How would we know if we’re using it too much?”
Co-creating rules, for example limiting AI use for homework until after a first attempt alone, or avoiding AI for emotional questions, helps children internalize boundaries as part of their own judgment rather than external constraints. This is particularly important with AI, because these systems are designed to keep engagement going.
There is also a regulatory dimension to consider. Many current protections rely on parental consent, but consent alone is not enough when children (and even parents) cannot fully understand how these systems work. Family rules can therefore act as a practical layer of governance at home.
Parents can, for instance, establish simple habits: always asking “Would I make the same decision without AI?”, keeping certain spaces “AI-free” such as meals or bedtime, or discussing together surprising or uncomfortable AI responses. These small, concrete practices translate abstract ideas like safety, privacy, and autonomy into everyday behaviors that children can actually apply and remember.
From a cybersecurity perspective, to truly and effectively protect your child, the rules must also be technically verifiable and simple to implement. For example, it’s possible to block certain websites and platforms using a DNS service like Cloudflare’s “1.1.1.1” which is known for blocking adult content or malware. The application is free and available on all platforms. You can also opt for other solutions, such as Apple, Google or Microsoft parental controls. Since I have a daughter who has an iPhone at home, I recently set it up through Family Sharing. This freedom that Apple Parental Control offers allows me to give her my permission if she wants to install an application or if she wants to buy content, such as an e-book. I can also block the device at any time by setting a screen time limit or being notified if she has arrived at school or at home. For example, I can also adjust the parental control according to her age.
What parents can do:
The rules must have children and adolescents participation so that they feel fully implicated in the process. It should also be remembered that this type of device should under no circumstances be installed as a monitoring device using real-time location. In my case, for example, I regularly consult parental control and look for any suspicious behavior. For example, I identify risky behavior, then I decide to talk about it with my wife and finally with my daughter.
In addition, treat the rules as something you revisit regularly. Children grow quickly, and so does technology. You can, for example, schedule short family check-ins every few weeks or month to ask: “Are our rules still working?” or “Have you used AI in a new way recently?”
More examples of what you can do as a family:
- Define clear moments of permission, for example when AI is allowed and when it is not, rather than only focusing on limits.
- Agree on simple signals your child or teen can use if something feels uncomfortable (a word or phrase to ask for help without explaining everything immediately).
- Encourage your child to show you how they use AI, instead of only asking them to follow rules.
- Create a shared understanding of what “good use” looks like, not only what is forbidden.
- Gradually give more responsibility, adjusting rules as your child or teen demonstrates good judgment.
- How can you protect your child’s digital identity (personal data, photos, privacy)?
Protecting a child’s digital identity today goes beyond traditional personal data. AI systems can infer sensitive aspects such as preferences, emotional states, or behavioral tendencies from seemingly trivial interactions. This means that even when data appears minimal, the inferences generated can be deeply personal and long-lasting. From an ethical standpoint, this raises questions about dignity and the right to develop one’s identity without constant observation and profiling.
From a cybersecurity standpoint, you can protect your data in different ways to minimize and partition it. It’s necessary to minimize the association with the full name, hide faces in photos, not display the location in publications, use different email addresses for each service, use pseudonyms, and regularly review the privacy settings of the applications. I also recommend that parents search their child’s digital footprint to inventory the public content, such as typing their child’s first name or email address into Google and reviewing the results. If, for example, your child appears in search results, I also recommend requesting deletion or delisting. This approach aims to protect the child. I also recommend that you look at sites like ‘’Have I Been Pwned’’ to see if an account of their child has not leaked on the dark web.
What parents can do:
In addition to limiting what is shared on the Internet, it is important to think about how information accumulates over time. Even small pieces of data can build a detailed profile. Try to regularly ask: “What could someone learn about my child or teen from this content I want to post?”
Additionally, you can add better controls to the tools your family is using. For instance, in Gmail, you can add a word after a “+” in your address to identify the origin of the message and better detect misuse of your address.
More examples of what you can do as a family:
- Encourage your child and yourself to think before sharing by asking: “Would you be comfortable if this stayed online for years?”.
- Avoid linking accounts together (for example, using the same login across multiple platforms), which makes tracking easier.
- Check which apps or games your child uses and whether they require creating a profile or answering personal questions.
- Disable automatic backups or sharing features that upload photos, conversations, or voice recordings without clear visibility.
- Teach your child that not everything needs to be documented or shared, especially everyday moments.
- How can you help your children develop critical thinking when faced with AI-generated content?
From a cybersecurity perspective, it’s a question of how we must strengthen the child’s rational autonomy to distinguish the true from the false in a digital space where fake content is often more valued than reliable content.
In cybersecurity, users are often advised to take some time to think (a few minutes) before acting. This allows us to appeal to our rationality rather than to our emotions, which can mislead us.
However, it must be explained that different signals can help identify AI-generated content. We can mention in particular: inconsistencies in hands or shadows, imperfect lip-sync, overly smooth voices, very generic text or text without an identifiable source. Even if no AI is perfect, their accumulation can be a negative signal.
It is advisable to encourage the child to perform reverse image searches, go to fact-checking websites. or identify AI-generated content labels. The goal isn’t to transform the child into a cyberspy, but to give him the psychological and practical tools necessary to exercise his critical thinking about digital content.
In addition, critical thinking in the age of AI is not just about knowing what is true or false, it is also about helping children understand how AI works. AI does not “know” things like a person does. It generates answers based on patterns it has learned from large amounts of data.
A simple way to explain this to a child is: “AI is very good at sounding right, but it doesn’t always understand what it is saying.” It is also important to help children notice how AI can feel convincing.
Sometimes the answers sound very confident, friendly, or even caring, even when they are not fully correct. You can encourage your child to ask simple questions like: “How do we know this is true?” or “Should we check this somewhere else?” The goal is not to make them distrust everything, but to help them stay curious, question what they see, and not take every answer at face value, especially when it sounds very sure of itself.
What parents can do:
Rather than simply warning children about fake content, try to make critical thinking part of everyday conversations. When your child shows you something from AI, avoid giving the answer immediately. Instead, ask: “What do you think about this?” or “What makes this answer convincing?” This helps them build their own judgment.
More examples of what you can do as a family:
- Encourage your child to slow down before reacting, especially to surprising or emotional content.
- Turn it into a habit to check one piece of information together, showing them how to verify without making it complicated.
- Play small “spot the clue” games, asking them to identify what looks strange or too perfect in a video, image, or text.
- Ask your child to compare two different answers to the same question and notice differences.
- Remind them that confidence does not mean accuracy, especially when something sounds very certain.
- When your child or teen starts creating content with AI, what new risks should you be aware of?
When children use AI to create content, they are no longer just using a tool, they are also shaping what goes online. For parents, the goal is not to stop children or teens from using these tools, but to guide them. You can encourage your child to use AI as a helper, not a replacement. For example: ask them to start an idea on their own before using AI, or to explain in their own words what they created. It is also useful to talk about responsibility, reminding them that anything they share online, even if generated with AI, is still linked to them. The key is to help children stay creative and confident in their own voice, while using AI as support, not as a substitute.
From a cybsersecurity perspective, access to AI is now extremely easy, even without advanced technical skills. This means it can also be misused. For example, AI can be used to generate harmful messages, manipulate images, or support forms of school harassment.
A simple photo of your child could be altered using AI tools and turned into something inappropriate or misleading, which can cause serious emotional harm. For this reason, it is important for parents to help children understand both the possibilities and the risks of using AI.
Beyond image manipulation, it is important to understand that data shared with AI tools, especially free ones, may be reused to build profiles over time. This means that seemingly harmless information can contribute to a broader picture of your child’s life, preferences, or family environment. For example, if a child shares personal details, such as family issues or emotions, this information may influence the type of content, recommendations, or even advertisements that appear later. In some cases, data may also be exposed, reused, or accessed by third parties. A simple photo shared online can be taken, modified, or used out of context. For parents, protecting a child’s digital identity means being mindful not only of what is shared publicly, but also of what is entered into digital tools.
.
What parents can do:
When your child uses AI to create something, try to stay involved in the process, not only the result. Ask simple questions like: “What part did you create?” and “What did the AI help you with?” This helps them stay aware of their role and not lose their own voice.
More examples of what you can do as a family:
- Encourage your child to keep a clear boundary between personal life and AI, avoiding sharing private family matters, photos, or sensitive information.
- Remind them to think before uploading or generating images, especially of themselves or others.
- Talk about how AI can be misused by others (for example, changing or misinterpreting images) so they understand the risks.
- Ask them to review content before sharing, checking if it could be misunderstood or used in a harmful way.
- Set a simple rule that creative work should always include their own input, not only AI-generated content.
- Do you know what kind of relationship is your child building with AI?
AI is not just a tool your child uses, it is something they may start to relate to. Because these systems are designed to respond quickly, patiently, and without judgment, they can feel easier to talk to than people. Over time, a child may begin to trust AI for advice, reassurance, or even emotional support. This does not happen suddenly, but gradually, through repeated interactions that feel helpful and comforting.
The challenge is that AI can simulate understanding without truly knowing your child. It can reinforce their thoughts, agree with them, or give confident answers, even when the situation requires nuance or human context. This can influence how children see themselves, make decisions, or choose who to turn to. Helping children build a healthy relationship with AI means ensuring it remains a support tool, not a substitute for human connection, reflection, or guidance.
What parents can do:
Try to understand not just how often your child uses AI, but why. Ask simple questions like: “When do you like to use it?” or “What do you ask it that you wouldn’t ask someone else?” This can reveal whether AI is becoming a source of comfort, advice, or validation.
More examples of what you can do as a family:
- Encourage your child to talk to people first when something is emotional or important.
- Gently remind them that AI does not truly understand them, even if it sounds like it does.
- Notice if your child starts preferring AI over friends, family, or teachers.
- Create moments where decisions are made without AI input, especially for personal matters.
- Reinforce that real relationships involve disagreement, nuance, and unpredictability, which AI cannot replicate.
Useful Sources for Families:
Adolescents & Anthropomorphic AI: Rethinking Design for Wellbeing (Everyone.AI, 2026): https://everyone.ai/research/
CNIL. (2026). IA conversationnelle et santé mentale des jeunes : résultats de l’enquête européenne.
https://www.cnil.fr/fr/ia-conversationnelle-et-sante-mentale-des-jeunes-resultats-de-lenquete-europeenne
Common Sense Media. (2025). Talking to AI: How teens use AI companions.
Youth, AI, and the Relationships That Shape Them (The Rithm Project, 2026) :
https://drive.google.com/file/d/1oGI3uxseNdfIc2Pms19TsKBQOB7wnJhV/view
How Gen Z Uses Gen AI and Why It Worries Them (Harvard Business Review, 2026) :
https://hbr.org/2026/01/how-gen-z-uses-gen-ai-and-why-it-worries-them
Talk, Trust and Trade-Offs: How and Why Teens Use AI Companions (Common Sense Media, 2025) :
Helping Parents Understand the Conversations Their Teens Are Having With AI (Meta, 2026) : https://about.fb.com/news/2026/04/helping-parents-understand-conversations-their-teens-are-having-with-ai/
SAIFCA :
https://www.safeaiforchildren.org/ai-risks-to-children-full-guide
Techs :
To learn more about Cloudflare DNS : https://one.one.one.one/family
To learn more about Apple’s parental control : https://www.apple.com/family-sharing
To learn more about Gmail Plus address Trick : https://unanswered.io/guide/gmail-plus-address-trick
To learn more about Have I Been Pwned : https://haveibeenpwned.com/