AI Parenting Goes Mainstream Across Europe, and the Risks Are Being Ignored

From Entertainment to Medical Guidance

What began as a novelty has become routine. Parents are asking chatbots to manage behavioural issues, interpret symptoms when children fall ill, and in some cases act as extended digital caregivers for hours at a stretch. A 2024 study found that a proportion of parents actively trust ChatGPT over qualified health professionals, believing the information it generates over expert medical advice. That finding should alarm anyone working in European healthcare.

The age at which children are being exposed to these tools is also falling. By 2023, roughly 30 per cent of parents with school-aged children were already using ChatGPT regularly; that figure has grown since. In the EU alone, where smartphone and broadband penetration rates are among the highest in the world, the conditions for rapid adoption are firmly in place.

The Safety Problem Nobody Wants to Name

The core danger is not that AI gives wrong answers occasionally. It is that AI gives wrong answers confidently, and parents under time pressure do not always stop to verify them. Chatbots are known to hallucinate, generating plausible-sounding but factually incorrect information, and their tendency toward sycophantic responses means they rarely push back on a user's assumptions. For a tired parent at midnight trying to decide whether a child's fever warrants a trip to accident and emergency, that is a genuinely hazardous dynamic.

There are more acute risks too. Friend-style AI companions have been linked in multiple studies to the intensification of emotional distress among teenagers, and in documented cases to suicidal ideation. The EU's own AI Act, which entered into force on 1 August 2024, classifies systems that interact directly with vulnerable users, including minors, as requiring specific transparency and safety obligations. Whether the current generation of consumer chatbots meets those obligations is, at best, an open question.

Andrea Jelinek, former chair of the European Data Protection Board, has consistently argued that children deserve categorical protection under EU data law, not merely a checkbox acknowledgement in a terms-of-service document. Her position is unambiguous: consent mechanisms designed for adults are structurally inadequate for protecting minors in AI environments.

Usage Patterns Tell a Complicated Story

The data on how parents and teenagers actually use these tools reveals a significant gap between parental comfort and teenage practice. Parents are broadly comfortable with AI for entertainment and homework assistance, less so for emotional support or medical queries. Teenagers, predictably, are using the tools more extensively across every category.

• Entertainment and games: Parent comfort above 70 per cent; teen usage above 80 per cent
• Homework assistance: Parent comfort around 45 per cent; teen usage around 65 per cent
• Emotional support: Parent comfort at roughly 18 per cent; teen usage around 35 per cent
• Medical advice: Parent comfort as low as 12 per cent; teen usage still around 25 per cent

That last figure is the one that should concentrate minds at health ministries and hospital trusts. One in four teenagers already using AI for medical queries represents a serious public health exposure, not a hypothetical future risk.

Privacy: The Under-Discussed Vulnerability

Data privacy compounds the safety problem. When parents type their child's symptoms, full name, school, or family circumstances into a commercial chatbot, that information enters a corporate data pipeline with limited transparency about retention, processing, or third-party sharing. Under GDPR, personal data relating to minors carries heightened protection, but enforcement of those rules in the context of AI-generated conversations remains patchy.

The specific risks are worth stating plainly rather than burying in a paragraph:

• Corporate retention of children's personal data and potential for secondary commercial use
• Vulnerability to data breaches exposing intimate family details
• Lack of transparency about how children's conversational data is processed and shared
• Potential access by malicious actors to sensitive family information
• Long-term consequences for a child's digital privacy as they mature into adulthood

Lena Radauer, a senior researcher at the Austrian Institute of Technology focusing on digital rights and AI governance, has noted that the gap between what GDPR promises and what AI companies deliver in practice remains uncomfortably wide, particularly for products marketed to general consumers rather than specialist enterprise users.

Using AI Responsibly in the Parenting Context

None of this means AI is useless to parents. Used as a first-pass research tool, it can surface relevant questions to bring to a GP, suggest age-appropriate activities, or help a non-native speaker draft a letter to a school. The problem is the substitution dynamic: when chatbot convenience displaces the consultation with a qualified professional rather than preparing for it.

That framing, AI as a starting point rather than a final authority, is the only defensible position for anything touching children's health or emotional development. A chatbot that gives an 80 per cent accurate answer is fine for drafting a party invitation. It is not fine for determining whether a child's rash needs antibiotics.

What Responsible AI Parenting Looks Like in Practice

For parents who want to engage with these tools without exposing their families to unnecessary risk, a few principles hold regardless of which platform they are using:

1. Treat AI output as a prompt for further enquiry, not a conclusion
2. Never input identifying personal information, medical histories, or sensitive family details into a commercial chatbot
3. For any health concern, consult a qualified professional; use AI only for preliminary background research if at all
4. Explain to children that they are interacting with a machine, not a friend or counsellor, and that its responses are not inherently trustworthy
5. Watch for signs of over-reliance: preferring AI interaction to human contact, accepting AI advice without question, or distress when access is restricted

The EU AI Act's provisions on transparency and high-risk systems will, over time, impose minimum standards on how these tools operate around minors. But enforcement will lag adoption by years. In the interim, the burden falls on parents, paediatricians, and schools to build the kind of AI literacy that the technology companies have little commercial incentive to promote themselves.