EU KIDS Act
Proposal for a REGULATION OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL EU KIDS ACT - 'EU Keeping Internet Digital Spaces Accountable and Trustworthy'
European Union
RAI-EU-NA-CO20266-2026COM(2026) 681 final
EU KIDS Act is Proposed in European Union as of 20 Oct 2026, according to eur-lex.europa.eu.
RegulationFundamental RightsData Protection and PrivacyGovernance and OversightProposed EU rules protect minors online through account-age limits, safer service design, AI safeguards and private age verification.
Summary
The proposed EU KIDS Act harmonises protection for minors on social networks, video platforms, games, AI companions, chatbots and app stores. It introduces account-age restrictions, safety-by-design duties and privacy-preserving age assurance, with supervision linked to the DSA, AI Act and GDPR. It is not enacted legislation.
Full article
Read full text ↗Overview
The EU KIDS Act proposes harmonised protection for minors online under Article 114 of the Treaty on the Functioning of the European Union. The Commission adopted the proposal, COM(2026) 681 final, on 17 September 2026 under ordinary legislative procedure 2026/0286(COD). Commission adoption is not enactment by Parliament and Council. Its 43 articles across nine chapters combine access restrictions, protective design, parental-responsibility verification, age assurance, national support and enforcement. The Commission feedback period closes on 3 December 2026 at midnight Brussels time.
Article 2 covers social networking, video-sharing, app stores, online games, operating systems, AI companions and general conversational chatbots accessible to minors. Digital-service providers fall within scope when serving recipients established or located in the Union, regardless of provider establishment; relevant AI systems are covered when placed on the Union market or put into service there. Exclusions encompass specified nonprofit reference resources, institution-operated educational services, certain open-source development platforms, exclusively scientific research systems and exclusively public-authority systems. The open-source exclusion does not apply where the platform itself is an AI system within scope. Duties vary by category: operating systems do not inherit every safety-by-design obligation.
Definitions
Article 3 defines a minor or child as a person under 18 and a guardian as a person holding parental responsibility under national law. An AI companion provides sustained, personalised interaction or companionship simulating or facilitating a social, emotional or interpersonal relationship. A general conversational chatbot directly interacts with users and assists across multiple domains and tasks. Specialised conversational functionality limited to a service, task or predefined functions is excluded, including customer-service, operational, technical-support, transactional, educational, information-retrieval and industrial applications. Relational functionality and broad conversational capability therefore provide distinct routes into scope.
Age assurance determines, estimates or verifies age and excludes self-declaration. Age verification establishes an age or threshold with high certainty using identification documents or other reliable verified identification sources. An age signal is a token, credential, attribute or information demonstrating an age, threshold or range. A proof-of-age attestation electronically authenticates that its holder meets an age condition. An EU age-verification solution requires public-authority certification under the EU Age Verification Scheme and inclusion in the Commission list following Member State notification. Other terms are incorporated from the DSA, DMA, AI Act and electronic-identification framework.
Governance and Institutional Framework
Article 34 applies DSA Chapter IV supervision to social networking services, video-sharing platforms, video gaming platforms and app stores. AI Act Chapter IX governs AI companions and general conversational chatbots, with the proposal's cross-reference to Chapters II, III and IV. Member States must empower their designated DSA and AI Act authorities. Standalone video games require a designated authority with DSA Article 51 powers; the authority at the provider's main establishment has exclusive supervisory and enforcement competence. Data protection authorities separately oversee personal-data processing, particularly under Articles 27–29. National decisions must not contradict Commission decisions under the regulation.
Applicable Commission DSA and AI Act powers extend to supervision, investigation, enforcement and monitoring. Article 36 establishes annual fees for each covered service or system under Commission supervision. Aggregate fees cover the preceding year's supervisory costs, including personnel, the age-verification scheme, enforcement and expertise. Each provider's fee is capped at 0.03% of preceding-year worldwide annual net income, rather than turnover; delegated acts establish methodology and payment arrangements. A Child Safety Committee assists with implementing acts. Commission and national authorities develop expertise and incident-response cooperation.
Account Access and AI Safeguards
Article 6 prohibits accounts below age 15 on social networking and video-sharing services with specified risk features: public real-time transmission, interaction beyond existing connections, profiling-based recommendations, recommendations outside subscriptions or connections, or uninterrupted-consumption and re-engagement designs. A derogation permits guardian-created limited accounts for minors described as above 13 and below 15. Guardian tools must remain activated, daily use cannot exceed one hour, and contacts require guardian pre-approval and limits. Article 7 permits specially designed video-sharing services to give children below 13 access only through a guardian's account, with no access below three. An impact assessment, age-adapted features, harmful-content exclusions and tightly restricted personalisation are required. These are conditional permissions, not access entitlements.
Article 14 requires AI companions and general conversational chatbots to prevent addictive designs and simulated relationships likely to create emotional dependency. It applies Article 9's general prohibition and time-limiting measures, Article 11 safe settings and Article 13 transaction safeguards. Information or analysis from prior interactions must not inform subsequent interactions by default, except where necessary for safety or those settings. Below-13 access must be enabled and controlled through guardian tools. State-of-the-art pre-market evaluation and testing must address health, safety, fundamental rights and physical, mental and emotional development, followed by safeguards. Post-market harm assessment and serious-incident response are required, except for micro and small enterprises. Embedded AI in social, video-sharing or gaming services must not auto-activate, receive prominent display or encourage minors' use; easy opt-out must remain available.
Safety Design and Service Implementation
Article 8 makes child-protective design the default, including accountless access, unless compliant age assurance establishes adulthood. Social and video-sharing services must prevent compulsive use, introduce meaningful interruptions and protect school time and core sleep hours. Recommenders must prioritise explicit preferences, disable implicit engagement-based recommendations by default, avoid off-service personal data and assess quality, safety and mental-health outcomes. Minors receive preference-reset tools and a nonprofiling option without manipulative steering. Geolocation, tracking, microphone, camera, contact synchronisation, account recommendations and push notifications are off by default. Changes require the specified age condition and informed explicit consent; location settings reset after sessions. Harmful social-comparison and image-distortion features are unavailable to minors.
Contact safeguards include prior approval, agreement to group entry, easy anonymous blocking, limited visibility, protection of contact details and restrictions on downloading or screenshotting minors' information or content. Transactions require clear real-time explanations and local-currency equivalents for purchasable virtual currency. Excessive-spending designs and variable reward systems are prohibited. Article 15 applies particular safeguards to games, including protection from diversion to risky external contacts and support for compliant user-created games. App stores must rate every application, publish rating methodology, prevent inappropriate access or purchases, respect higher legal age limits and admit certified EU age-verification solutions.
Articles 18–20 require understandable controls, durable effects from feedback, child-friendly reporting and support, and guardian tools respecting minors' privacy and agency. Tools must be age-tailored, secure and effective, and minors must be informed when they are used. Article 24 requires relevant non-EU providers to designate a representative or extend an existing qualifying DSA or AI Act mandate to support enforcement.
Monitoring, Audits and Evaluation
Article 5 requires designated very large social networking and video-sharing platforms to notify detailed compliance plans within four months of designation; already designated providers must do so within 30 days of application. Providers finance independent audits drawing on child rights, paediatrics, child psychiatry, developmental science, age assurance, interface and recommender design, and data protection and security expertise. Auditors receive relevant information and personnel access. Employees and contractors communicating with auditors must be protected from retaliation. Providers have 15 days to comment on a draft report. The final report reaches the Commission and provider simultaneously within two months of receipt of the plan, followed without undue delay by a public nonconfidential summary.
A Commission decision identifying shortcomings triggers a corrective plan within 30 days. Each measure must have an implementation period of no more than 60 days, verified by auditors. Neither an audit nor Commission inaction establishes compliance or restricts enforcement. Annual reporting and auditing integrate into DSA risk assessments and audits; Article 22 also requires VLOPs to monitor, test and evaluate safety measures. Article 14 retains separate AI pre-market and post-market duties, with its micro- and small-enterprise monitoring exception. Commission-assessed codes may demonstrate relevant AI or gaming compliance but do not confer blanket exemptions. Article 42 proposes a first review on the bracketed date 31 August 2030, assessing effectiveness, scope, internal-market outcomes and freedom of expression, with reports every four years.
Penalties, Complaints and Redress
Article 34 incorporates DSA enforcement for specified platform categories rather than a universal penalty schedule. Intentional or negligent breaches of covered AI-companion and chatbot duties attract AI Act Article 99 administrative fines of up to 6% of the undertaking's preceding-year worldwide annual turnover. Member States establish standalone video-game penalties using DSA Article 52. Data protection authorities may impose GDPR Article 83 fines within their competence, up to its Article 83(5) ceiling. These penalties are distinct from the supervisory fee's 0.03% net-income cap. The proposal does not establish a standalone general damages tariff.
Article 35 provides expedited Commission proceedings for designated VLOPs and AI systems under exclusive Commission supervision. The Commission must endeavour to communicate preliminary findings within the bracketed 30 working days and adopt a final decision within 90 working days after proceedings open; these are endeavour-based targets. Article 29 requires free electronic internal complaints about incorrect age-assurance outcomes. Article 19 requires prioritised handling of minors' reports and information on procedure and redress. Article 21 allows qualifying nonprofits to represent minors and guardians and complain to the competent authority where the recipient or user is located or established, or to the European AI Office for AI systems within its exclusive competence. Existing redress and representative-action rights remain preserved.
Relationship to Other EU Instruments
The proposal specifies and complements the Digital Services Act and complements the Artificial Intelligence Act. Article 2(6) deems qualifying social networking, video-sharing and video gaming platforms compliant with DSA Article 28(1) for covered matters when they comply with this regulation, including relevant AI functionalities. This is limited alignment, not a general DSA exemption. Under Article 8(3), only Chapter III Sections III, V and VI apply to AI companions or chatbots deployed as social-networking or video-sharing functionalities; AI-specific obligations remain relevant.
Article 2 preserves audiovisual-media, consumer-protection, product-safety, personal-data and other listed Union rules. Article 6 access restrictions operate alongside consent-based processing rules in Article 8 of the GDPR, not as their replacement. DMA definitions and relevant guardian-tool interoperability conditions are incorporated, while age attestations connect to the European Digital Identity Wallet framework. Article 41 proposes adding the instrument to Annex I of Directive (EU) 2020/1828 on representative actions. The explanatory memorandum places the proposal alongside BIK+, cyberbullying support and planned audiovisual-media and consumer-law work; those future initiatives remain distinct.
Age Assurance and International Alignment
Establishment-neutral scope harmonises cross-border provision to the Union. Article 31 establishes Union-wide mutual recognition of certified, listed EU age-verification solutions and proof-of-age attestations, not general recognition of third-country credentials. Article 29 requires a prescribed certified third-party solution for Article 6 account restrictions. Alternative assurance methods may serve designated safety-by-design and app-store purposes if they satisfy Articles 27 and 28. Article 32 allows the stated high-confidence exception for existing-account checks.
Age assurance must be accurate, reliable, secure, robust, nonintrusive and nondiscriminatory; every measure must use zero-knowledge proof. It must not identify, locate, track, target, advertise to or profile recipients. Strict minimisation and purpose limitation apply, with a narrow permission to retain an account-level threshold signal. Operating systems holding compliant age signals must enable sharing when required for compliance following user consent. Member States must provide accessible, free age-verification and parental-responsibility means, including alternatives for vulnerable families and people without standard documentation. The explanatory memorandum connects harmonisation to the Union's International Digital Strategy and global child-safety cooperation, without establishing a treaty or general external recognition regime.
Implementation Timeline
| Milestone | Timing | Effect |
|---|---|---|
| Commission adoption | 17 September 2026 | COM(2026) 681 final; procedure 2026/0286(COD); proposal, not enactment. |
| Feedback closes | 3 December 2026 | Midnight Brussels time. |
| Proposed entry into force | Twentieth day after Official Journal publication | Article 5 applies from entry into force. |
| General application | Six months after entry into force | General proposed obligations apply. |
| Already designated VLOP plans | Within 30 days of application | Article 5 compliance-plan notification. |
| Existing-account checks | By six months after application | Article 6(4) checks and disabling of covered underage or age-unestablished accounts. |
| National support and expedited proceedings | 12 months after entry into force | Articles 33 and 35 apply; national strategies communicated by this point. |
| Age-rating and gaming codes | One year after application; assessment at 42 months after entry into force | Drawing-up target followed by adequacy assessment. |
| First Commission review | 31 August 2030 | Bracketed proposed Article 42 date; subsequent reports every four years. |
Sources and References
Requirements for a company
What an organisation has to do under EU KIDS Act, at a glance. Not legal advice — the table below gives the provision and deadline for each item.
Not yet in force (Proposed). These requirements apply once the instrument takes effect and may change before then.
Must do
0Nothing in this category.
Must not do
0Nothing in this category.
Should do
9- Prepare to test AI risks to children's health, safety, rights and development before release, and implement safeguards.Providers of AI companions and general conversational chatbots available to minors in the EU
- Prepare to assess harms after release and respond to serious incidents.EU-facing AI companion and general chatbot providers, excluding micro and small enterprises
- Prepare to enforce applicable guardian controls, including limited-account contact approval, one-hour daily limits and guardian-controlled access for children under 13.Covered social networking, video-sharing, AI companion and general chatbot providers offering permitted child access
- Prepare to apply category-specific child-safe defaults, usage breaks, safe recommendations and disabled sensitive features unless compliant age assurance establishes adulthood.Covered EU-facing service providers, according to their service category
- Prepare to use certified third-party verification where required, zero-knowledge proofs, minimal data retention and free electronic complaints about incorrect results.Covered providers performing age assurance for EU users
- Prepare to provide understandable controls, effective feedback, prioritised child reporting and secure guardian tools that respect minors' privacy and explain their use.Covered EU-facing providers subject to child-control, reporting and guardian-tool requirements
- +3 more in the table below
Should not do
6- Prepare to prevent addictive AI designs and simulated relationships likely to create emotional dependency.Providers of AI companions and general conversational chatbots available to minors in the EU
- Prepare to disable reuse of previous interactions by default, except where needed for safety or required settings.Providers of AI companions and general conversational chatbots available to minors in the EU
- Prepare to prevent automatic activation, prominent display and encouragement of minors' embedded-AI use, while providing an easy opt-out.Providers embedding AI in EU-facing social networking, video-sharing or gaming services
- Prepare to block covered under-15 accounts except permitted guardian-created accounts, and check existing accounts before disabling prohibited or age-unverified accounts.EU-facing social networking and video-sharing providers with Article 6 risk features
- Prepare to exclude harmful social-comparison and image-distortion features for minors, and prohibit excessive-spending designs and variable reward systems.Covered EU-facing providers subject to the relevant feature and transaction safeguards
- Prepare to prevent age-assurance data from identifying, locating, tracking, targeting, advertising to or profiling users.Covered providers and organisations processing age-assurance data
Who must do what
The obligations under EU KIDS Act, most serious first. Not legal advice — verify against the official text before relying on it.
| # | Who | Requirement | By when | Where | Severity |
|---|---|---|---|---|---|
| 1 | Providers of AI companions and general conversational chatbots available to minors in the EU | Prepare to prevent addictive AI designs and simulated relationships likely to create emotional dependency. “prevent addictive designs and simulated relationships likely to create emotional dependency” | If enacted: six months after entry into force | Article 14 | Recommended |
| 2 | Providers of AI companions and general conversational chatbots available to minors in the EU | Prepare to disable reuse of previous interactions by default, except where needed for safety or required settings. “Information or analysis from prior interactions must not inform subsequent interactions by default” | If enacted: six months after entry into force | Article 14 | Recommended |
| 3 | Providers of AI companions and general conversational chatbots available to minors in the EU | Prepare to test AI risks to children's health, safety, rights and development before release, and implement safeguards. “State-of-the-art pre-market evaluation and testing must address health, safety, fundamental rights and physical, mental and emotional development” | If enacted: before placing on the EU market or putting into service | Article 14 | Recommended |
| 4 | EU-facing AI companion and general chatbot providers, excluding micro and small enterprises | Prepare to assess harms after release and respond to serious incidents. “Post-market harm assessment and serious-incident response are required, except for micro and small enterprises.” | If enacted: ongoing after release once applicable | Article 14 | Recommended |
| 5 | Providers embedding AI in EU-facing social networking, video-sharing or gaming services | Prepare to prevent automatic activation, prominent display and encouragement of minors' embedded-AI use, while providing an easy opt-out. “Embedded AI in social, video-sharing or gaming services must not auto-activate, receive prominent display or encourage minors' use” | If enacted: six months after entry into force | Article 14 | Recommended |
| 6 | EU-facing social networking and video-sharing providers with Article 6 risk features | Prepare to block covered under-15 accounts except permitted guardian-created accounts, and check existing accounts before disabling prohibited or age-unverified accounts. “Article 6 prohibits accounts below age 15 on social networking and video-sharing services with specified risk features” | If enacted: new-account restrictions at general application; existing-account checks within six months after application | Article 6 | Recommended |
| 7 | Covered social networking, video-sharing, AI companion and general chatbot providers offering permitted child access | Prepare to enforce applicable guardian controls, including limited-account contact approval, one-hour daily limits and guardian-controlled access for children under 13. “Below-13 access must be enabled and controlled through guardian tools.” | If enacted: before enabling the relevant child access | Articles 6, 7 and 14 | Recommended |
| 8 | Covered EU-facing service providers, according to their service category | Prepare to apply category-specific child-safe defaults, usage breaks, safe recommendations and disabled sensitive features unless compliant age assurance establishes adulthood. “Article 8 makes child-protective design the default, including accountless access, unless compliant age assurance establishes adulthood.” | If enacted: six months after entry into force | Articles 8–14 | Recommended |
| 9 | Covered EU-facing providers subject to the relevant feature and transaction safeguards | Prepare to exclude harmful social-comparison and image-distortion features for minors, and prohibit excessive-spending designs and variable reward systems. “Excessive-spending designs and variable reward systems are prohibited.” | If enacted: six months after entry into force | Articles 8–13 | Recommended |
| 10 | Covered providers performing age assurance for EU users | Prepare to use certified third-party verification where required, zero-knowledge proofs, minimal data retention and free electronic complaints about incorrect results. “Age assurance must be accurate, reliable, secure, robust, nonintrusive and nondiscriminatory; every measure must use zero-knowledge proof.” | If enacted: when applicable age-assurance requirements begin | Articles 27–29 | Recommended |
| 11 | Covered providers and organisations processing age-assurance data | Prepare to prevent age-assurance data from identifying, locating, tracking, targeting, advertising to or profiling users. “It must not identify, locate, track, target, advertise to or profile recipients.” | If enacted: whenever performing age assurance | Articles 27–29 | Recommended |
| 12 | Covered EU-facing providers subject to child-control, reporting and guardian-tool requirements | Prepare to provide understandable controls, effective feedback, prioritised child reporting and secure guardian tools that respect minors' privacy and explain their use. “Tools must be age-tailored, secure and effective, and minors must be informed when they are used.” | If enacted: six months after entry into force | Articles 18–20 | Recommended |
| 13 | EU-facing game, gaming-platform and app-store providers, according to their category | Prepare to protect gaming users from risky external contacts, rate applications, prevent inappropriate access and purchases, and accept certified EU age-verification solutions. “App stores must rate every application, publish rating methodology, prevent inappropriate access or purchases, respect higher legal age limits” | If enacted: six months after entry into force | Article 15; app-store article unspecified | Recommended |
| 14 | Designated very large EU-facing social networking and video-sharing platforms | Prepare to submit compliance plans, fund independent audits, protect auditor contacts from retaliation, publish summaries and implement verified corrections. “A Commission decision identifying shortcomings triggers a corrective plan within 30 days.” | If enacted: plans within four months of designation, or within 30 days of application for already designated providers | Articles 5 and 22 | Recommended |
| 15 | Relevant non-EU providers and covered providers under Commission supervision, respectively | Prepare to appoint an EU representative where required and pay annual supervisory fees when under Commission supervision. “Article 36 establishes annual fees for each covered service or system under Commission supervision.” | If enacted: representation at general application; supervisory fees annually under implementing payment arrangements | Articles 24 and 36 | Recommended |
More AI regulation in European Union
AI regulation in European Union: full overview
- Digital Omnibus on AI
- European Union - AI Coordination Plan
- EU 2026 Cybersecurity Resilience Package
- European Union - Digital Regulation Simplification (2025)
- EU Dual-Use Export Control Regulation
- EU AI Act
- European Union - Product Liability Reform (2024/2853)
- European Union - AI System Definition Guidelines
© Regulations.AI — created on 20 Oct 2026 using Gemini 3.6 Flash · reviewed against official sources on 20 Oct 2026