The proposal impacts social media and video sharing platforms (VSPs), certain online gaming, AI companion and conversational chatbots and app stores

In the EU President’s State of the Union speech, 16 September 2026, the EU KIDS ACT (Keeping Internet Digital Spaces Accountable and Trustworthy) was announced. The draft proposal has now been released. We set out a quick view below. 

The proposal:

• Sets out EU wide rules to ensure the same level of protection for children across the EU.
• For social media or VSPs with ‘risk’ features, no under 13s are permitted to have accounts and only safe, guardian set-up accounts are permitted for 13-15s.
• Is not limited to social media and VSPs, ‘safety by design’ rules and commitments will apply to certain online gaming and AI companion/chat services too. App stores are also subject to obligations on age ratings and age gating. 

Impact

The proposed EU KIDS ACT will make a considerable change to the EU’s current overall online safety framework, setting a mandatory safety baseline. 

The requirements apply to all users of the service (including unregistered) and the safety baseline is a default; it is only permissible to derogate from that position when it is established (by using age assurance) that the user is an adult. This is commercially significant and creates a broad user-experience impact for the services affected. 

The proposal builds on the Digital Services Act (DSA) and AI Act but more services are brought into scope than under those regulations. For social media and VSPs that are Very Large Online Platforms under the DSA, compliance obligations are increased. Those platforms must produce and independently audit a compliance plan describing how the platform complies with the EU KIDS ACT age restrictions and safety by design rules and report annually, assessing effectiveness and reporting as part of their annual DSA risk assessments. 

What is the proposal?

Scope

The scope is “Social Media+”, covering online social networking services (“social media”) and VSPs as well as online games, AI conversational chatbots and companions which have specific risk features and putting requirements on app stores. 

Structure

The proposal is for a “gradual and differentiated approach” which depends on the service type, the presence of specific ‘risk factors’ and the age of the child user. 

For social media and VSPs with ‘risk’ features (such as live streaming or profile-based recommender systems):

• users under 13 are not permitted to have accounts  
• users aged 13-15 can have accounts set up by guardians with mandatory safety features (always activated guardian controls, time restrictions and contact controls)
• users aged 15-18 can create autonomous accounts (but the accounts will not be “adult” versions because safety by design requirements will apply)

There is a caveat for VSPs which are designed for under 13s. They can enable a child limited access to the service through the guardian’s own account, subject to various restrictions.  

All services in scope must embed “safety by design” and are required, as a general obligation, to ensure a high level of privacy, safety and security for minors, which echoes DSA. In addition, there are general obligations to provide agency to minors (such as accessible and understandable information and user controls), child-friendly reporting and tools for guardians.

There are then specific minimum requirements depending on the service type. 

For all social media and VSPs the requirements are:

• no addictive design features like infinite scrolling
• regulated recommender systems, including providing user choice over the parameters and recommendations that are not only engagement based 
• safe account default settings, such as no geolocation or tracking features 
• no unwanted contact from other users, including controls on connecting and blocking ability
• transparency of and controls over economic transactions online

For AI companions and general conversational chatbots there is an additional obligation which echoes the AI Act, requiring a “high level of protection of the health, safety, fundamental rights and well-being and development of minors”. The mandatory requirements are tailored to these types of AI services and include avoiding design features that simulate interpersonal relationships creating emotional dependencies, for instance. However, also included is a broad requirement to evaluate and test for risks to the high level of protection and to implement safeguards, as well as post-market monitoring and mitigation. 

For providers of online games the requirements include ensuring excessive use of the game by minors is not encouraged.

For app stores the requirements include to put in place an age-rating system for all apps and the stores must make apps inaccessible to the underage. 

Conclusion

The proposal comes at a time when many EU Member States are trying to implement national child user social media bans and running into difficulties and legal challenge. For instance, the French Courts recently blocked a social media ban on the basis of its impact on children’s rights to free expression and privacy concerns. The KIDS ACT proposal provides a unified pan-EU approach, providing a quicker route to the new safety baseline, extending across many service types and preventing fragmentation risk from Member States having different rules.  

The EU proposal can expect to be scrutinised, potentially with respect to the balance of the various fundamental human rights it will achieve. The differentiated approach, with tiered age brackets, will presumably be said to tune that balance more than outright bans. The EU proposal contrasts to the UK proposal for a U16 social media ban on certain services and safety features announced earlier in the year. Details on the UK proposal are expected later this year. Potentially the EU’s differentiated approach is not helpful to the UK’s intention to impose more sweeping measures and need to justify the balance of fundamental rights that will achieve; particularly when risk management under Online Safety Act is already in place. We can expect to see more legal disputes as these proposals progress.  

Meanwhile, Australia is proposing to add a broader duties-based safety framework, having already implemented a U16 social media ban. In the United States, significant litigations and investigations are shaping the risk landscape for platforms. Against this global back-drop, although the EU KIDS ACT seeks to harmonise the position across the EU, it does have prescriptive safety requirements, necessitate new account types and set the user experience to a child-safe one unless a user is proven to be an adult. That will require not only potentially significant product change but, with the global landscape fragmented, such specificity may be challenging for platforms seeking to implement child safety and age assurance as consistently as possible on a global basis. 

There is more to come. As the EU President highlighted, “it is not only minors who are at risk. Addictive design, for example, is harming everyone. This is why we need a wider framework, the Digital Fairness Act.” The EU’s Digital Fairness Act proposal is due in the Autumn.

Want to know more?

If you would like to discuss the EU’s proposal and the issues in different regions with our global digital regulation experts, please do get in touch.