The European Commission will present an EU Kids Act on Thursday, September 17, 2026, that would make 15 the age at which children across the bloc can open their own accounts on higher-risk social media and video-sharing platforms, according to a draft Commission Communication and reporting by POLITICO and Euractiv published on September 14.
In Short
The European Commission has written a plan that would stop children under 15 from opening their own social media accounts on risky platforms, letting younger kids use them only through accounts their parents set up and control. It matters because the same plan would force apps like TikTok and Instagram to check ages at sign-up, switch off features such as endless scrolling for young users, and in some cases get Brussels’ approval before launching new features. The draft is not law yet, and the three documents behind this story do not fully agree on how the youngest age groups would be treated. Governments and the European Parliament still have to negotiate the text after Thursday.
What the draft sets out
The document at the centre of the story is a Communication from the Commission to the European Parliament and the Council, titled “An EU approach to online child safety” and carrying the reference COM(2026) 680/2. It is marked sensitive until adoption, its Strasbourg date line still reads “XXX”, and every page bears an Euractiv watermark. Section numbering is uneven in places, and a handful of typographical errors remain, which is consistent with a text still being finalised.
According to the draft, the EU Kids Act “puts forward a single, EU-wide set of rules providing equal level of protection for all children throughout the Union,” alongside a set of tools for parents and guardians. The Commission frames the aim as reducing fragmentation across member states and giving businesses legal certainty in the single market. A footnote specifies that the definition of parenthood will follow national legislation in each member state.
Scope is defined through a category the draft calls “Social Media+”. According to the Commission, this covers social media services, video-sharing platforms and online games that pose specific design risks for minors, plus AI chatbots and AI companions, which the text describes as virtual tools capable of giving mental health and personal development advice to minors. Services designed and operated for education or by public authorities are excluded, as are industrial AI and AI office products.
The proposal combines two instruments. One is an EU-wide minimum age. The other is a set of safe design requirements. According to the draft, the design rules apply to every service in scope, while the minimum age applies only to “risky” social media and video-sharing platforms. The text does not say how a platform would be classified as risky, a gap POLITICO also flagged when it reported that the limits apply to high-risk services without clarification of how that category would be determined.
Four age brackets
The draft sets out a graduated structure rather than a single cut-off.
For children aged 0 to below 3, there would be no access to social media or risky services. The Commission’s reasoning centres on relational risk: parents replacing direct interaction with screens, with potential effects on attention span, language, vision and socio-emotional development.
Between 3 and below 13, parents could open fully parent-controlled accounts on child-friendly services that meet strict safety standards appropriate to the child’s developmental stage. Those services would themselves need to fall within the regulation’s scope and comply with its safety-by-design criteria.
Between 13 and below 15, parents could set up introductory accounts on social media and video-sharing platforms subject to the minimum age. According to the draft, these accounts would carry strictly limited features, a maximum number of peer-only contacts, and limited screen time per day. Parents and guardians would monitor and control use through a dedicated guardian app. The draft does not specify the contact ceiling or the daily time limit.
From 15 to below 18, teenagers could open autonomous accounts without parental approval, provided the environment is safe by design.
Where the sources diverge
The three sources do not line up neatly on the lower brackets, and the differences matter for how the final text might read.
According to POLITICO, parents could open fully guardian-controlled accounts for children under 13 on services meeting strict safety standards. That report does not mention the separate prohibition for children under 3 found in the draft Communication.
Euractiv’s account points somewhere else. According to Euractiv, the Commission was considering a blanket ban on access for children under 13, combined with a requirement that 13- to 15-year-olds obtain parental consent. A Commission official told the outlet that the approach was still under development, and Euractiv described the age threshold as a key detail yet to be disclosed.
The draft Communication does not describe a blanket ban for all under-13s. It describes a ban only below age 3, and parent-controlled access between 3 and 13. Whether the version adopted on Thursday follows the draft, the Euractiv description, or neither, cannot be established from the material available. POLITICO itself noted that the text remained subject to change before formal presentation.
Safe design obligations
Beyond age thresholds, the draft lists design rules that would bind all services in scope. According to the Commission, they include:
- No addictive design features, with infinite scrolling, artificial notifications and certain reward features given as examples.
- Regulated recommender systems, including options to choose, tune and control recommended content, measures to avoid rabbit hole effects, and a requirement to go beyond engagement-based signals.
- Safe account settings that keep content private, disable risky settings by default and inform minors unambiguously.
- A block on unwanted contacts, under which unknown users could not initiate direct contact outside approved connections, explicit permission would be needed before a minor is added to a group, and blocking would be easy and anonymous.
- Easy reporting of harmful content and access to support tools.
- Effective and age-appropriate parental control tools.
Those rules would sit alongside tailored commitments for online games, AI chatbots and AI companions. The Commission describes this strand as co-regulatory, built on industry-led self-regulation. Software application stores would also face requirements, according to the draft, building on current self-regulatory practice: consistent age ratings, and apps such as online games made inaccessible to minors below the rated age.
The draft is explicit about what current law lacks. According to the Commission, the existing EU framework does not specify a minimum age for platforms with risky features, contains no outright ban on endless scrolling, excess notifications or harmful recommender systems, and lacks legal certainty on both age verification systems and the legal responsibility of technology companies for the design of their services. It also acknowledges that the dark patterns and addictive design features found in digital products are not addressed with sufficiently strong harmonisation today.
Age verification at sign-up
The mechanics of checking age are among the most detailed parts of the text. According to the draft, providers of social media services and video-sharing platforms would be obliged to perform age verification whenever a user opens a new account. That verification would rely on the EU age verification tool or on other solutions provided by public authorities offering an equivalent level of accuracy, reliability, robustness, privacy, security, non-discrimination and non-intrusiveness.
Existing accounts would be treated differently. The Commission describes proportionate checks rather than blanket re-verification, citing as examples accounts created many years earlier or accounts linked to a credit card held by the account owner.
Games get their own track. According to the draft, an obligatory age recommendation would be required for every game, with strict enforcement through age verification in a software application store when the game is downloaded. A separate Euractiv item listed alongside its report, dated September 14, carried the headline that gamers will have to prove their age under the act.
The draft also contains a pre-market check. To prevent social media services from reaching children before their safety design has been fully verified by public authorities, the proposal would require verification of such services before they are made available to minors. The text does not state which authority would carry out that verification or on what timetable.
Brussels has been building the underlying tooling for more than a year. The Commission published technical specifications for an EU-wide system on July 14, 2025, and later adopted a non-binding recommendation pushing member states to deploy privacy-preserving verification apps by December 31, 2026. The Kids Act draft would turn verification at account creation from an option into an obligation for the services in scope, a shift with direct consequences for how age assurance is designed across the industry.
Enforcement modelled on the DSA and AI Act
According to the draft, the EU Kids Act builds on the enforcement architecture of the Digital Services Act and the AI Act, and places the burden of proof for compliance on operators. POLITICO reported the same design, noting that those laws centralise enforcement of the highest-risk companies at the Commission while national regulators oversee smaller firms.
The most consequential mechanism for large platforms is a prior-approval step. According to the Commission, very large online platforms and very large online search engines would have to submit a compliance plan to the Commission, which would assess it within 30 days. A new service, feature or functionality could only be rolled out after a positive opinion. The draft does not say whether this applies to every new feature these companies ship, or only to those accessible to minors. That distinction will determine whether the provision touches the product roadmaps of the largest platforms in a limited way or across the board.
The Act would also create a supervisory fee to fund the Commission’s EU-wide enforcement and supervision work, according to both the draft and POLITICO. Neither document gives a rate or a cap. Complaints related to minors, or coming from minors, would receive priority within the enforcement structure. A review clause would allow the framework to be updated as technology changes.
Penalty levels are absent from the material supplied. Euractiv listed a separate exclusive alongside its report headlined as saying von der Leyen would fine technology companies under the act, but the body of that piece is not among the documents reviewed here.
The DSA enforcement record gives some indication of scale. The Commission’s preliminary findings against TikTok on February 6, 2026 and against Meta on July 10, 2026 both targeted infinite scroll, autoplay, push notifications and recommender systems, with potential fines of up to 6% of global annual turnover if confirmed. Both cases remain at the preliminary stage, and both rest on the systemic risk duties that apply only to designated services. The draft Communication cites these proceedings as evidence that enforcement is already a priority, and separately refers to action against X over Grok’s generation and dissemination of child sexual abuse material. That case followed the Commission’s January 26, 2026 decision to extend its X proceedings to Grok.
The evidence base Brussels cites
The draft opens with figures. According to a June 2026 Eurobarometer cited by the Commission, teenagers spend an average of 4.5 hours online on school days, rising to 6.1 hours on weekend days. Those who started using social media before age 10 report 7.5 hours of weekend screen time, compared with 5.7 hours for those who started after 14.
Other numbers follow. According to the Commission, one in three teenagers report feeling stressed or socially excluded and affected by concentration problems. One in four children encounters harmful and distressing content online. One in four experiences online sexual solicitation at some point before 18, a figure the draft sources to Childlight. Pressure to share intimate pictures affects one in ten teenagers. The draft also cites the OECD’s PISA results for 2025 as confirming that digital distraction affects learning and contributes to declining basic skills in the EU.
The Commission links these harms directly to business models. “Platforms’ business models, designed to maximise the time spent online, result in a massive scale of online engagement and may exploit the vulnerabilities of children,” according to the draft.
Its conclusion is blunter. “Parents, not algorithms, should be raising Europe’s children,” the draft states, adding that technology companies bear the primary responsibility for making their products safe. Elsewhere it says the Union’s approach “needs to limit the access of tech companies to our children, and not the other way around.”
The political track began with the 2025 State of the Union address, after which a Special Panel was tasked with advising von der Leyen on a European approach, according to the draft. POLITICO reported that the panel of health and technology experts recommended restricting access to what it called “social media+”, including AI companions, earlier this year.
A crowded national landscape
Brussels is not acting in a vacuum. According to a footnote in the draft, legislation is being prepared or negotiated in 17 member states. POLITICO named five that are pushing ahead with national laws to restrict children’s access: Austria, Denmark, France, Greece and Spain. It also reported that the Commission’s move follows months of pressure from government leaders including French President Emmanuel Macron.
France illustrates the legal risk. French lawmakers approved an under-15 ban on July 21, 2026, after removing the clause that would have compelled platforms to build age verification systems. The Constitutional Council then struck the law down on August 14, 2026, eighteen days before it was due to apply. According to Euractiv, the French government submitted a reworked national proposal on Monday, September 14. The draft Communication cites the French decision in a footnote, acknowledging that recent developments demonstrate “the legal complexity surrounding the issue” and the need for a proportionate approach that protects fundamental rights.
Outside the EU, the models differ. Australia’s under-16 ban, which took effect on December 10, 2025, names nine platforms and allows no parental consent exemption. The United Kingdom set out an under-16 ban on June 15, 2026, with a Spring 2027 target. The EU draft is structurally softer at the top end, setting the autonomous-account age at 15 and allowing supervised access below it.
The draft also points to the United States, describing a recent Meta settlement as evidence that service providers are increasingly regarded as responsible for children’s safety. Its footnote links to a Virginia Attorney General release on a settlement with Meta. PPC Land reported that a federal judge entered a consent judgment on August 26, 2026 imposing a default two-hour daily limit on teen use of Instagram and Facebook, with payments priced at up to about $17.2 billion over ten years.
Why the marketing industry is watching
For advertisers, agencies and publishers, the draft touches supply, targeting signals and product timelines at once.
Verification at account creation would change what platforms know about their users. Under the DSA, Article 28 already bars profiling-based advertising to users a provider knows with reasonable certainty to be minors. Mandatory age checks would convert that “reasonable certainty” from a probabilistic estimate into a verified attribute for new accounts. The practical effect has precedent. Reddit disabled ad personalisation for 13- to 15-year-olds across the EU from June 24, 2026 when its own age checks began. In Australia, Meta removed access from 756,000 under-16 accounts in seven months, changing the composition of the audiences that feed its recommender and lookalike systems.
Recommender systems and reach
The design rules may matter more than the age gate for most campaigns. A requirement that recommender systems “go beyond engagement-based signals” for services in scope goes to the core of how feed-based platforms distribute organic content and, indirectly, how they sell attention. The draft does not say whether the rules apply only to minors’ accounts or to the whole service. Nor does it define which signals would qualify. PPC Land’s coverage of the July 2026 Meta findingsnoted that advertisers buy Instagram and Facebook inventory largely because those recommender systems are effective at maximising time spent. If the same logic is written into primary legislation, the question for buyers is how much inventory, and at what engagement rates, survives.
The 30-day prior approval for new features adds another variable. Ad products are features. If the provision applies broadly, launch calendars at the largest platforms could lengthen in the EU relative to other markets.
Influencers, games and AI chatbots
Several adjacent files widen the scope. According to the draft, the forthcoming Digital Fairness Act will update consumer rules on manipulative business-to-consumer practices, unfair pricing, problematic influencer marketing, and digital contracts and subscriptions. The advertising industry has already contested that file: an IAB Europe coalition objected to the consultation’s scope in July 2025, arguing that existing law covers the ground. The Commission also intends to revise the Audiovisual Media Services Directive to specify the categories of harmful content video-sharing platforms must protect minors from, and to clarify how EU audiovisual rules apply to influencers as content providers.
Games publishers face the app store age-rating and download-time verification requirements. Mobile game user acquisition, a significant line of app-install spending, runs through those same stores.
AI chatbots are explicitly inside “Social Media+”. That places conversational assistants carrying advertising alongside social networks in the scope of the design rules. ChatGPT’s designation under the DSA took effect at the end of August 2026, which would also place it in the tier subject to the 30-day compliance plan procedure, should the draft survive negotiation in its current form.
Complementary measures
The Communication packages the Act with a list of other initiatives. According to the Commission, it will propose a Digital Fairness Act, propose a revision of the Consumer Protection Cooperation Regulation, propose a revision of the Audiovisual Media Services Directive, present an Action Plan on protecting children from crime, and present an Education package.
The CPC revision would consider improving coordination between national consumer authorities in cross-border cases and centralising investigation and enforcement powers at EU level in specific cases. The draft cites earlier CPC work, including an investigation into Star Stable Online over direct purchase exhortations aimed at children and unclear in-game currency information, and ongoing coordinated action against Temu and Shein over dark patterns, fake discounts and other unfair practices.
The Education package would mainstream AI literacy through an EU-OECD framework for primary and secondary schools and address the supply of European digital education infrastructure. The EU co-funded Safer Internet Centres would provide tailored guidance on the new rules for children, guardians, educators, schools and regulators. Under Horizon Europe, the 2026 work programme covers the effect of digital tools on educational outcomes and mental health, with projects starting in 2027. Erasmus+ would continue funding offline activities.
On health, the draft recalls a 2023 Communication on mental health that included 20 flagship initiatives and almost EUR 1.3 billion in funding. With UNICEF, the Commission is developing a prevention toolkit due for World Mental Health Day in October 2026. It is also working with the World Health Organization on capacity-building in mental health systems.
Internationally, the EU has worked with Australia, Brazil, Canada, Singapore, the United Kingdom and Japan, and helped shape the G7 Common Set of Principles on minors online, which the draft says was subsequently endorsed by leaders from Brazil, Egypt, Kenya, India and South Korea.
Monitoring would draw on the EU Children’s Participation Platform, the Better Internet for Kids+ strategy, DSA work and the EU Centre for Algorithmic Transparency, with a specific brief to track whether the minimum age framework needs adjusting.
What happens next
The calendar is compressed. According to Euractiv, all 27 commissioners were due to exchange views on the draft on Tuesday, September 15, with adoption expected this week by written procedure. Von der Leyen is expected to address the structure of what Euractiv described as a social media “delay” in her State of the Union speech in Strasbourg on Wednesday, September 16. Commission spokesperson Olof Gil confirmed on Monday that the act would come on Thursday, according to POLITICO, and Euractiv reported that von der Leyen and Executive Vice-President Henna Virkkunen will present it together.
Thursday is a starting point, not an end. POLITICO noted that the text will have to be negotiated with national governments and members of the European Parliament. Euractiv reported separately that Ireland, which holds the rotating Council presidency, will prioritise negotiations on EU-wide rules. The Commission declined to comment on the plans, according to POLITICO.
Several questions remain open in the documents reviewed. How “risky” platforms are identified. What the contact and screen-time caps for introductory accounts would be. Whether the 30-day feature approval covers every launch. What the supervisory fee and penalties amount to. And whether an EU instrument can withstand the kind of proportionality challenge that sank the French law in August. The draft’s own language concedes that last point is not settled.
Timeline
- November 29, 2024: Australian Parliament passes its under-16 social media law
- July 14, 2025: European Commission publishes technical specifications for an EU-wide age verification system
- July 14, 2025: IAB Europe coalition challenges the scope of the Digital Fairness Act consultation
- September 2025: State of the Union address leads to the creation of a Special Panel on child safety online
- December 10, 2025: Australia’s under-16 ban takes effect
- January 26, 2026: Commission extends DSA proceedings against X to Grok
- February 6, 2026: Commission issues preliminary DSA findings against TikTok over addictive design
- April 29, 2026: Commission issues preliminary findings against Meta over under-13 access and adopts a recommendation on age verification apps
- June 2026: Eurobarometer on screen time and youth mental health, cited in the draft Communication
- June 15, 2026: UK government sets out an under-16 social media ban targeting Spring 2027
- June 24, 2026: Reddit disables ad personalisation for EU teen accounts as age checks begin
- July 10, 2026: Commission issues preliminary findings on addictive design at Instagram and Facebook
- July 21, 2026: French Parliament approves an under-15 ban
- August 13, 2026: Meta reports blocking 756,000 under-16 accounts in Australia
- August 14, 2026: France’s Constitutional Council strikes down the under-15 ban
- August 26, 2026: Federal consent judgment imposes a two-hour daily teen limit on Instagram and Facebook
- Late August 2026: ChatGPT falls under DSA risk rules
- September 14, 2026: POLITICO reports the EU KIDS Act plans; Commission spokesperson Olof Gil confirms Thursday timing; France submits a reworked national proposal
- September 15, 2026: Commissioners due to exchange views on the draft
- September 16, 2026: State of the Union speech in Strasbourg
- September 17, 2026: Scheduled presentation of the EU Kids Act by Ursula von der Leyen and Henna Virkkunen
- October 2026: UNICEF and Commission prevention toolkit due on World Mental Health Day
- 2027: Horizon Europe projects on digital tools and youth mental health due to start
Summary
Who: The European Commission, with President Ursula von der Leyen and Executive Vice-President Henna Virkkunen presenting the proposal. Platforms affected include social media services, video-sharing platforms, online games, AI chatbots and AI companions, app stores, and the very large online platforms and search engines subject to prior approval. Reporting came from POLITICO and Euractiv, with Commission spokesperson Olof Gil confirming the timing.
What: A proposed EU Kids Act that would set 15 as the minimum age for autonomous accounts on risky social media and video-sharing platforms, with parent-controlled access below that age and no access below 3 according to the draft. It would add safe design rules, including no addictive features and recommender systems that go beyond engagement signals, mandatory age verification at account creation, a 30-day Commission review of compliance plans before large platforms roll out new features, and a supervisory fee. The sources disagree on treatment of under-13s.
When: Reported on September 14, 2026. Commissioners were due to discuss the draft on September 15, von der Leyen speaks in Strasbourg on September 16, and the proposal is scheduled for presentation on Thursday, September 17, 2026, followed by negotiations with the Council and the European Parliament.
Where: All 27 EU member states, replacing a patchwork in which 17 member states are preparing or negotiating national legislation, according to the draft.
Why: The Commission argues existing law sets no minimum age, does not ban addictive features, and lacks legal certainty on age verification, citing Eurobarometer data showing teenagers spend 4.5 hours online on school days and 6.1 hours on weekends. France’s national ban was struck down in August 2026, strengthening the case for a harmonised EU instrument that can withstand proportionality review.
