Meta, OpenAI face crackdown under new EU Kids Act

The European Commission is set to propose the EU Kids Act, a digital safety law that would ban social media access for children under 13 and require parental supervision for users aged 13 to 15. The draft targets platforms including social networks, video-sharing sites, online games, and AI chatbots. Commission President Ursula von der Leyen formally unveiled the plan during her State of the European Union address yesterday. 

The proposal comes amid pressure from EU member states, including France, Denmark, and Spain, which are advancing or considering national social media bans. The Act also follows Australia’s decision to ban social media for under-16s, reflecting a global trend toward tighter digital protections for minors.

Tech giants like Meta, OpenAI, and Anthropic will face significant compliance costs under the Act – and fines up to 6% of their revenue if found in breach — particularly for AI chatbots and social media features. The supervisory fee and potential fines could reshape how platforms design services for minors, with many expected to overhaul age-gating systems and content moderation.

Age-based access tiers and restrictions 

The EU Kids Act introduces a phased approach to digital access, dubbed a 'digital age ladder.' Children under 3 are barred entirely from high-risk services. For ages 3 to 13, access is limited to curated, child-friendly services with full parental supervision. Users aged 13 to 14 may open 'introductory' accounts on mainstream platforms, but only with explicit parental consent, daily screen-time limits, and restricted features such as limited stranger contact. 

Minors aged 15 and above can access 'safe by design' versions of social media, video-sharing, gaming, and AI chatbot services without parental oversight. The framework explicitly excludes educational tools, public services, and AI systems for office or industrial use. 

Safety-by-design and addictive feature bans

Platforms must implement safety-by-design standards to minimize harm to minors. This includes banning infinite scrolling, autoplaying videos, push-notification triggers, and engagement-driven reward loops like 'streaks' or 'likes.' Recommender systems must avoid 'rabbit holes' and prioritize non-personalized content to prevent algorithmic amplification of harmful material. 

Children’s accounts must default to private, and unknown users cannot initiate contact with minors. The Act also requires platforms to disable risky settings by default and provide age-appropriate parental control tools. These measures build on the EU expert panel’s recommendations, which urged shifting the burden of proof onto tech companies to demonstrate compliance.

Mandatory age verification 

The Act mandates age verification at account creation for social media and video-sharing platforms, and before download for games. The EU plans to deploy an official open-source age verification app using zero-knowledge cryptographic proofs, allowing users to confirm age without disclosing identity or exact birthdate. 

However, legal experts warn of circumvention risks, such as teens using siblings’ credentials. Platforms classified as 'Very Large Online Platforms' under the Digital Services Act must submit compliance plans for Commission approval 30 days before launching new features affecting children. Non-compliance carries fines of up to 6% of global annual turnover, with a supervisory fee of up to 0.03% of global sales to fund enforcement.

Restrictions on AI chatbots 

The EU Kids Act explicitly extends child-protection rules to AI chatbots and companions, treating them as high-risk services alongside social media and gaming. Platforms must ensure AI systems avoid designs that foster unhealthy emotional attachment or addictive behaviors. The Act requires AI chatbots to adhere to the same age restrictions, safety-by-design principles, and parental control mandates as other covered services. 

This marks the first time the EU has wrapped consumer AI into a unified child-protection regime, signaling a broader regulatory shift. The proposal aligns with the AI Act and Digital Services Act, creating a layered enforcement framework.

Setting a precedent 

The draft Act is still subject to negotiation between the European Parliament and Council, with the final shape likely to evolve. Legal and technical challenges remain, including the feasibility of age verification and the risk of circumvention. The Commission’s proposal formalizes the EU’s ambition to set a global precedent in digital child protection, but its success hinges on overcoming these hurdles.