A child safety law is usually sold as a morality play. For product teams, it becomes a sign-up screen, a permissions model, a data retention decision, and several meetings with counsel that somehow all happen on Friday. Reuters reported that the European Union is set to propose restrictions on under-15s using social media and AI chatbots. The practical question is not whether Brussels is worried about children online. It is whether your onboarding flow can tell a 14-year-old from a 15-year-old without collecting more personal data than the privacy team can defend.

What Reuters says is coming

According to Reuters, the European Union will propose a ban restricting under-15s from social media, video sharing platforms, AI chatbots, and online games. SRN News, carrying the Reuters report, says the proposal is part of an EU Kids Act that European Commission President Ursula von der Leyen and EU tech chief Henna Virkkunen will present on Thursday, with von der Leyen possibly announcing details in her State of the Union address on Wednesday.

This is still a proposal, not an enforceable rule, which is the part LinkedIn will forget by lunchtime. SRN News reports that a European Commission document framed the policy aim this way: "The Union’s approach should allow minors to benefit from the huge potential of digital services, while protecting them from abuse and exploitation. It needs to limit the access of tech companies to our children, and not the other way around." That sentence is doing a lot of work. In compliance terms, it points toward restricting company access to minors by default, not merely adding a parental control page three clicks deep.

Who is in scope, at least on the draft facts

Reuters identifies the affected categories as social media, video sharing platforms, AI chatbots, and online games. SRN News notes that concerns have focused on major services including Meta’s Facebook and Instagram, TikTok, Google’s YouTube, and ChatGPT. That list is not just a naming exercise. It tells consumer AI teams that chatbot access is being treated alongside feeds, video, and games, not as a harmless text box because it speaks in complete sentences.

Ground News reports that the draft would extend safety-by-design rules to video games and AI chatbots, require companies to verify users’ ages, and require tools for children to report harmful content. It also reports that companies would pay supervisory fees to fund enforcement while avoiding addictive designs and harmful feeds. Translate that into product work: age assurance at account creation, minor-specific defaults, reporting pathways that work for children, and design reviews for feeds or engagement loops that currently optimize for time spent. If your internal answer is that age lives in a marketing database and trust and safety will handle the rest, the lawyers are not going to enjoy that architecture diagram.

The age tiers are the product requirement

The Edge Malaysia, citing a Bloomberg News report on a policy document, says the EU would establish a minimum age of 15 to open an autonomous account, meaning an account without parental supervision. Ground News reports a tiered model: under-3s receive no access, children aged 3 to 12 require full parental control for child-friendly services, users aged 13 to 14 may request limited introductory accounts, and those 15 and older manage their own profiles. That is not a single age gate. It is four account states pretending to be a policy slogan.

For builders, the redesign starts with account flows. A service would need to decide whether it can offer no access, child-friendly access, supervised access, restricted introductory access, and autonomous access without creating loopholes between them. AI chatbot teams have an extra problem: access control is not enough if the same model behavior, memory settings, personalization, and content filters apply to every age tier. The useful work now is to map which features belong in each age state, then document why.

What is not law yet, and what should move anyway

Ground News says the Commission must negotiate the draft with EU countries and the European Parliament in the coming months before it becomes law. The Edge Malaysia also notes that the document is a draft that might change before publication. So no, the law does not yet require an immediate EU-wide shutdown of every under-15 account. It does mean teams that serve minors in Europe should stop treating age assurance as a future procurement exercise.

The safest near-term move is not to guess the final text. It is to make your system adaptable enough that a 13-year-old, a 14-year-old, and a 15-year-old do not all fall into the same bucket. Build the account taxonomy, define supervised versus autonomous permissions, test age verification vendors against privacy requirements, and separate chatbot access controls from general account status. If the proposal narrows, that work still improves governance. If it expands, you will be glad your minor experience is not a pile of feature flags named after a panic meeting.

Sources - EU is set to propose ban on social media and AI chatbots for under-15s

Sources