Quick Answer
The EU Kids Act would require AI chatbots and companions to be off by default for minors, under a European Commission proposal published September 17, 2026. The proposal also targets emotionally dependent behavior and wider child-safety settings, but it is not law yet. Families and platforms should watch the EU legislative process rather than treat the requirements as currently enforceable.
Key Takeaways
- The EU Kids Act proposal covers AI chatbots, AI companions, social media, online games, and video-sharing platforms used by minors.
- AI chatbots and companions would be turned off by default for users under 18.
- Chatbots could not behave in ways that make children emotionally dependent on them.
- Children under 13 would be barred from social media services, while users aged 13 and 14 could receive limited mini accounts.
- The proposal requires approval from EU governments and the European Parliament before it can become law.
What would the EU Kids Act require from AI chatbots?
The EU Kids Act would require AI chatbots and AI companions to be turned off by default for minors. The European Commission proposed the measure on September 17, 2026, as part of a broader package covering online services used by children, including social media, video-sharing platforms, online games, AI companions, and AI chatbots. The European Commission’s announcement describes the proposal as an effort to make online services safer for children.
The default-off requirement matters because it would move the decision away from a child-facing setting and toward a platform-level safety control. A minor could not simply encounter an AI companion as a standard feature unless the service had a permitted path to enable access. The proposal does not yet establish final technical requirements for every product, so companies do not have a current legal obligation to change chatbot settings.
For US readers, the immediate effect is informational rather than practical. The EU Kids Act is an EU proposal, but major AI services often operate across regions, and age-assurance changes developed for Europe can influence how platforms structure youth accounts elsewhere. Readers tracking broader platform safeguards may also want to understand how teen and adult age groups are being introduced on other online services.
Which AI behaviors would the EU Kids Act restrict for minors?
The EU Kids Act would restrict AI chatbot behavior that makes children emotionally dependent on a service. The Commission proposal specifically identifies AI companions and chatbots as services that need stronger protections for minors, rather than treating conversational AI as an ordinary messaging or search feature.
The proposed restrictions are important because AI companions can present themselves as attentive, responsive, and personally available. The draft described by WIRED would prohibit chatbots from simulating emotions or interpersonal relationships with users under 18. The same reporting states that chatbots would be prohibited from retaining memory between a minor’s separate chats. The reported draft details indicate that the policy would address both the content of an AI interaction and the continuity that can make an interaction feel like an ongoing relationship.
The memory restriction would limit a chatbot’s ability to build a persistent profile from one conversation to the next for a minor. That could reduce the sense that a chatbot knows a child personally, although it would not eliminate all risks from a single conversation. Parents should still discuss what personal information children should not share with any chatbot, including names, locations, school details, financial information, and private family matters.
How would default-off AI chatbot settings work for minors?
Default-off AI chatbot settings would mean an AI companion or chatbot feature starts disabled for a minor rather than active as part of a standard account. The EU Kids Act proposal would apply under-18 safety obligations to AI companions and chatbots, including this default-off approach.
The proposal also links chatbot controls to broader account protections. Minors’ profiles would be private by default, while geolocation, camera access, and microphone access would be disabled under the planned rules. Those settings matter because conversational AI tools can collect more sensitive context when a service can access a device’s location, images, audio, or profile information.
| Proposed protection | Who it would affect | Why it matters |
|---|---|---|
| AI chatbots off by default | Minors | AI companion access would not begin as a standard enabled feature. |
| Limits on emotional relationship simulation | Users under 18 | Chatbots could not be designed to simulate an interpersonal relationship with a minor. |
| No memory between separate chats | Users under 18 | Chatbots would have less ability to create continuity across a minor’s conversations. |
| Private profiles and disabled device permissions | Minors using covered services | Geolocation, camera, and microphone access would be disabled by default. |
The practical limitation is that the proposal does not yet specify every interface choice a platform must use. A service could need age assurance before deciding whether to show a chatbot feature, and the final law could alter the implementation details. Users should avoid assuming that an existing chatbot account has changed until a platform announces an actual policy update.
What age rules would the EU Kids Act set for social media?
The EU Kids Act would bar children under 13 from social media services under the Commission’s proposal. Users aged 13 and 14 could use parent- or guardian-managed mini accounts with limited features and a one-hour daily restriction, creating a separate access model for younger teenagers rather than a full adult-style account.
The age rules do not apply only to AI tools, but they provide the wider framework for services that may include chatbot features. A platform that offers social feeds, messaging, video, games, and generative AI within the same account would need to consider which protections apply to each part of the service. The policy direction resembles other efforts to separate youth experiences from adult accounts, including parental supervision and time limits on social platforms.
The one-hour restriction applies to the proposed mini accounts for 13- and 14-year-olds, not to every user under 18. Parents should not treat the proposal as a current parental-control setting, because no new EU Kids Act account type is available yet. Existing platform age rules remain the rules that govern an account until legislation passes and companies implement compliant products.
How would platforms verify a user’s age under the proposal?
The EU Kids Act proposal would require independent age-verification tools to support its age-based protections. The planned European Digital Identity Wallet is among the tools identified for that purpose, and platforms would need to reassess existing accounts within six months after the rules take effect, according to the proposal details.
Age assurance matters because default-off chatbot protections depend on a service accurately determining whether an account belongs to a minor. A platform cannot consistently apply under-18 restrictions if a child can easily enter an inaccurate birth date and keep an account classified as an adult account. Reuters’ breakdown of the proposal states that independent age-verification tools would be required.
The privacy question remains significant. The proposal points toward independent tools rather than leaving platforms to rely only on self-reported ages, but the final legislation would need to establish how those tools protect personal data and how services handle verification failures. Families should use official platform controls and avoid submitting identity documents to unofficial links or third-party services that claim to offer account verification.
What penalties could platforms face under the EU Kids Act?
The EU Kids Act could expose platforms to fines of up to 6% of worldwide annual turnover if the child-safety-by-design requirements become law in their reported form. The potential penalty is substantial because worldwide annual turnover measures a company’s broader revenue, not only the revenue from a specific youth product or EU feature.
The enforcement exposure is designed to make child-safety requirements a product and governance issue rather than a minor settings adjustment. A company offering an AI companion to minors could need to assess feature design, account classification, permission settings, memory behavior, and relationship-style prompts together. The proposed approach is more comprehensive than a single warning screen because several design choices can affect a child’s experience.
The 6% figure remains a proposed maximum, not a penalty that has been imposed. The European Parliament and EU governments can change the text, the scope, and enforcement mechanisms during negotiations. Companies should prepare for the policy direction, while users should avoid treating a possible future fine as proof that any particular AI service currently violates EU law.
When could the EU Kids Act become law?
The EU Kids Act is not law yet and has no confirmed implementation date. The European Commission proposed the measure on September 17, 2026, but EU governments and the European Parliament must approve legislation before the requirements can take effect.
The legislative process matters because the final text can change before adoption. The proposed chatbot rules, the social media age thresholds, the mini-account structure, age-verification requirements, and the penalty framework could all be revised through negotiations. The European Commission also presented the proposal in a September 17 press conference, which outlines the broader child-safety policy rationale.
The sensible response for parents is to review the actual settings available on services their children use now. Enable private accounts where available, review camera and microphone permissions, and discuss when a chatbot should not be used for emotional support or urgent personal problems. Contact a qualified health professional or emergency service when a child expresses immediate distress, self-harm concerns, abuse, or danger. An AI chatbot is not an appropriate replacement for emergency or clinical support.
What should parents and AI companies do before the rules change?
Parents should treat the EU Kids Act as a signal to review children’s current AI and social media access, not as a rule that platforms already enforce. Check whether a child’s account is private, whether location and device permissions are enabled, and whether the service offers a youth or supervised account option. A child should understand that an AI chatbot can generate responses without being a trusted person, therapist, or emergency resource.
AI companies should examine whether their youth experiences depend on persistent memory, relationship-style language, or device data access. The proposed requirements point toward safety-by-design controls that operate before a minor begins a conversation, rather than relying on a child to find and change advanced settings. Companies building AI features into devices and platforms may also face closer scrutiny of how conversational tools use account data.
Users should remain careful with AI products regardless of their age. Sensitive account details and recovery codes should never be pasted into a chatbot, particularly as AI-related scams continue to use familiar product names and interfaces to target users. TechJournal has covered fake AI trading agents that install wallet malware, which illustrates why account security and source verification remain necessary even when a service appears legitimate.
FAQ
What is the EU Kids Act?
The EU Kids Act is a European Commission proposal intended to make online services safer for minors. The proposal covers social media, video-sharing platforms, online games, AI companions, and AI chatbots used by children.
Would the EU Kids Act ban AI chatbots for minors?
No, the EU Kids Act proposal would not impose a complete ban on AI chatbots for minors. The proposal would require AI chatbots and companions to be turned off by default and would restrict emotionally dependent behavior.
Would AI chatbots be allowed to remember conversations with children?
No, the reported EU Kids Act draft would prohibit chatbots from retaining memory between a minor’s separate chats. The restriction would limit a chatbot’s ability to create an ongoing relationship-like experience across conversations.
Would the EU Kids Act affect children under 13?
Yes, the EU Kids Act proposal would bar children under 13 from social media services. Users aged 13 and 14 could use parent- or guardian-managed mini accounts with limited features and a one-hour daily restriction.
Is the EU Kids Act in effect now?
No, the EU Kids Act is not in effect now because it remains a European Commission proposal. EU governments and the European Parliament must approve the measure before any final requirements can take effect.