Social media now sit inside the daily lives of minors. Children under ten are already in the scroll, some of them for as much as eight hours a day. That is the picture the European Commission used to launch the KIDS Act: an age-graded protection system, a ban on social media for under-13s, and tighter duties for platforms.
“Too many children are exposed too early to the online world. They are not ready to move in an environment where bullying can follow you home and every mistake can be recorded forever,” Commission President Ursula von der Leyen said at a press conference in Strasbourg, presenting the bill she had previewed on Wednesday in her State of the Union speech.
She cited a recent Eurobarometer survey in which 92 percent of respondents want European action to protect children and young people online, and said about 70 percent of EU citizens worry about cyberbullying, harassment and grooming.
Von der Leyen blamed the platforms themselves: sophisticated technology built not for “the wellbeing and development of children,” but to capture attention, exploit behaviour and keep people clicking, reacting and scrolling — features that are part of the business model. “The goal is to keep children online and monetise their attention,” she said.
The Commission says the EU KIDS Act comes from a special expert group, consultations with young people, and a close look at countries that have already moved, including Australia.
The law, she said, “will set clear boundaries in the digital world, just as we do every day in the real world.” Under-13s would be barred from social media. Under-15s would not get a personal account — only mini-accounts created and supervised by parents or guardians, with limited functions and a cap of one hour a day. Between 15 and 18, safe design would become a legal duty for platforms.
“I want to be very clear: setting an age limit does not relieve tech companies of responsibility for the content on their platforms,” von der Leyen said. Platforms had “treated the mental health of our children too lightly.” They would now have to show they are safe, on the principle of safety by design.
What the proposal actually does
Worry about the health and safety effects of major platforms — Meta’s Facebook and Instagram, TikTok, Google’s YouTube, ChatGPT — has grown worldwide. Australia led the way last year with a social-media ban for under-16s. That pushed several European countries, France first among them, to consider similar limits. Some, especially in Scandinavia, argued the decision should stay with parents. To avoid a patchwork, the EU decided to act itself.
The core of the Commission proposal is an age-graded system plus new platform obligations.
Under 13: no social media. Video-sharing services designed for children would still be allowed, through guardian-managed accounts.
13 to 15: access only via “mini-accounts” created and run by parents or guardians, with age-appropriate services, limits on contacts and a one-hour daily cap.
15: minimum EU age to open an account independently.
Under 18: safety-by-design rules apply more widely.
Executive Vice-President Henna Virkkunen said parental control would not be only about time. Guardians could pre-approve new contacts and cap how many users a child may interact with. “Parents will be able to decide whether their children can access social media and, if so, they will do so under strict guardian supervision and with a set of safeguards.”
The second pillar is safety by design. It reaches beyond social networks to video-sharing platforms, online games, AI virtual assistants and chatbots used by under-18s. The Commission wants to ban features seen as harmful or compulsive: infinite scroll, reward systems, push notifications during sleep hours, and profiling-based recommendation feeds that steer minors toward harmful content. By default, minors’ profiles would be private; geolocation, camera and microphone would be off.
Virkkunen described a list of “do’s and don’ts” for digital services used by anyone under 18. “Strangers will not be able to message children. Minors will not be targets of personalised advertising or algorithms based on previous activity and, above all, addictive features will no longer be allowed.” The duties would cover social media, video platforms, online games, chatbots and “AI companions.”
AI gets a specific clamp. Virtual assistants and chatbots should be off by default for minors and should not simulate interpersonal relationships that create emotional dependence. “With the EU KIDS Act, they will no longer be able to behave like human friends in ways that make children emotionally dependent on them,” Virkkunen said. By default, systems could not reuse a child’s earlier conversations. They would have to be tested before launch for risks to children and monitored afterwards. Chatbots inside children’s games or apps could not switch on automatically or be pushed at minors, and they would have to be easy to turn off.
Platforms would also have to offer simple tools to block or mute other users, control time online, and change or reset recommendation systems.
Age verification would become mandatory for online services and app stores. One option is the European age-verification app, designed to prove age without storing ID documents or biometric data. Social networks and video platforms would check age when new accounts are opened; for existing ones they would have to estimate age from plausible indicators, such as the account creation date or credit-card data.
“Proving your age online should be simple and should reveal absolutely nothing else about the user except their age,” Virkkunen said. Parental tools should be “simple, interoperable and on by default,” with proof of parental responsibility in one click.
The proposal also flips the burden of proof. Large platforms would have to show that services, features and products aimed at minors are safe by design. Providers would submit a compliance plan to the Commission, reviewed by an independent auditor. If there are gaps, Brussels could demand corrective measures.
Virkkunen tied the new system to existing supervision. “We now have three years of Digital Services Act enforcement behind us and this year alone we have issued ten preliminary findings against the main platforms.” She named TikTok over addictive design and unsafe account settings, Meta, and four large pornographic platforms. The AI Act, she noted, already bans systems that manipulate people or exploit children’s vulnerabilities.
Enforcement would sit on DSA and AI Act structures, with a faster clock: if a breach looks likely, the Commission should finish investigations within 90 days. The text is still a legislative proposal, sent to the European Parliament and the Council, and can change in negotiation.
Age limits may not be enough
Italy’s Fondazione S.O.S. Telefono Azzurro welcomed the proposal but asked for stronger guarantees. “Europe finally recognises that children’s safety must be built into the design of digital services and guaranteed by companies before harm occurs,” said president Ernesto Caffo. “Age limits alone are not enough: we need measurable standards, truly transparent audits and a real link to listening and emergency services. No child should be shut out from asking for help.”
The foundation wants risk assessment of chatbots and AI companions to look not only at single answers but at the relationship that can form over time — emotional dependence and the risk of manipulation. “In cases of self-harm, suicide risk, grooming or sextortion, the technology must recognise the danger and guarantee an immediate handover to qualified people.”
In the Parliament–Council talks it wants: access to helplines and emergency services even without parental authorisation when that cannot be obtained safely; appeal procedures when age checks go wrong; and systems that can pass a crisis conversation to a human operator. It also wants independent audit results published, real sanctions when protection fails, dedicated channels for Trusted Flaggers, and structured — not merely consultative — participation of children and teenagers in the Commission’s monitoring.
Telefono Azzurro cited its own 2025 figures: 2,186 cases on the 1.96.96 line, 3,446 interventions via 114 Child Emergency, and 3,022 reports to the “Clicca e Segnala” hotline, more than 76 percent of them about child sexual abuse material. It also pointed to Internet Watch Foundation data: more than 8,000 AI-generated child sexual abuse images and videos identified in the same year.
The KIDS Act is now in the ordinary legislative procedure. Age gates will be the headline. The fight in Brussels will be over design duties, age checks, AI companions — and whether platforms have to prove safety before the harm is done.
The article in Italian by EUalive’s partner FocusEurope.it can be found here.
Caption: AI-generated photorealistic illustration of three 10–11-year-olds sitting side by side, each absorbed in a smartphone instead of playing together.
Updated: September 17, 2026 – 16:23
