EU Preliminarily Rules TikTok's Endless Scroll Violates DSA

What the EU Took Issue With in How TikTok Is Used

On February 6, 2026, the European Commission issued a preliminary finding that TikTok’s addictive design violates the Digital Services Act (DSA). It flagged infinite scroll, autoplay, push notifications, and a highly personalized recommendation system as the problems. This isn’t a final ruling on violation, nor has any fine been imposed yet.

What caught my attention is that the company is being held responsible not just for what videos users see, but for the design choices that keep them watching. In an earlier piece I wrote, Adult ADHD and the Living Environment Surge, I looked at how surrounding environments shape behavior; this time, the question is being asked of the companies that build the apps. I don’t mean to suggest that app use and ADHD are the same problem.

These features aren’t unique to TikTok. Other social media platforms also recommend what to watch next and use notifications to draw people back. So I found myself wondering what standard this ruling might set for other personalized services beyond TikTok.

The design that keeps you watching the next piece of content

Infinite scroll adds content every time you swipe down. It’s convenient—you never have to tap a “next page” button—but it also erases the natural stopping point where you’d otherwise feel you’d “finished” a page. Autoplay goes a step further, removing even the small act of choosing the next video.

Nir Eyal, who has written extensively about product design, explained the appeal of infinite scroll as tied to the anticipation of not knowing what interesting thing might come next. The effort required to decide whether to keep watching is minimal, while the opportunity to discover the next piece of content never stops presenting itself.

People often compare this to a slot machine, but scrolling a feed isn’t quite the same thing as gambling, and it’s not as if every user reacts identically either. To understand the design, what matters isn’t the metaphor but the concrete mechanics: when the next video starts, what you have to do to stop it, and what notifications arrive after you close the app.

The European Commission’s preliminary position is that TikTok failed to adequately assess the risks of these kinds of features. In particular, it found that TikTok didn’t properly account for indicators that could reveal excessive use—things like minors’ nighttime usage and how frequently the app gets opened. The Commission said this conclusion came after reviewing TikTok’s internal risk assessments and data, along with scientific research and expert opinions.

Is a Screen-Time Limit Button Enough?

The Commission concluded that existing time-management tools are too easy to disable, and that parental controls don’t sufficiently reduce risk because setting them up requires time and know-how. As examples of what needs to change, it pointed to things like halting infinite scroll after a set period, building in genuinely effective rest periods (including at night), and adjusting recommendation systems.

The issue I see here is the gap between “does a limiting feature exist” and “does it actually help people manage their usage.” Putting a feature in the settings menu is no guarantee that users understand it or actually use it. I think developers need to look not just at whether they’ve shipped a feature, but at how it’s actually being used in practice.

In a 2024 survey by the Pew Research Center in the US, about 60% of respondents aged 13 to 17 said they use TikTok daily, and 16% said they use it almost constantly. Since this is self-reported data from American teenagers, it can’t be directly translated into usage time or addiction-diagnosis rates for EU teens. Still, it’s useful data showing just how frequently this kind of service comes up in daily life.

The mere fact that someone uses a service for a long time doesn’t mean all of that time is harmful. What matters is whether users spent that long because they wanted to, or because they wanted to stop but found it hard to control themselves. Without making that distinction, I don’t think either approach — simply trying to raise or lower time-on-app — is good enough.

Other platforms are already under scrutiny too

The EU has been investigating Facebook and Instagram over child-protection issues since May 2024. That inquiry also covers concerns that algorithms and similar features might drive repetitive, compulsive use among young users. So this isn’t a case of TikTok being singled out for a feature unique to itself.

TikTok itself has a prior case on record too. In August 2024, the European Commission made legally binding TikTok’s commitment to permanently withdraw its TikTok Lite rewards program in the EU. This isn’t the first time product design meant to drive engagement has drawn regulatory attention.

If a DSA violation is ultimately confirmed, fines of up to 6% of global annual revenue can be imposed. The actual amount is set based on the nature and severity of the violation, among other factors. You can’t just multiply ByteDance’s target revenue by the cap percentage and present that as this case’s final fine.

I see this process as part of a broader shift: holding not just individual users but also the companies that build these services accountable for excessive use. If a company designs features that increase time spent on its platform, the reasoning goes, it also has to examine what effect those features have on users.

Oswarld’s Lens

Honestly, this ruling leaves me with two feelings at once: that it’s a necessary step, and that the scope of regulation could end up too broad.

I do think designs that make it hard for users to choose to stop watching, or that nudge people into behavior they didn’t intend, deserve regulation. Late-night notifications and repeated auto-play prompts should also be examined through the lens of user choice. Getting people to spend more time on a service isn’t always the same thing as a good product experience.

That said, whether restricting a specific service’s feature actually reduces how much people use it needs to be verified separately. Users might simply migrate to a similar app instead. To judge the real effect, we’d need to check not just whether usage of that particular app dropped, but whether the behaviors and discomforts that were actually harmful to users decreased too.

As someone who works with data, what concerns me more is personalization. The ability to remember a user’s context and surface the information they need matters for AI personal assistants too. Even when the same behavioral data is used, a design that helps users finish what they set out to do quickly and a design that keeps them engaged longer can have different purposes and different outcomes.

This preliminary ruling doesn’t mean all personalization algorithms are banned. But how far personalization will be allowed to go from here, and what harms companies will be required to mitigate, could still shape the design of AI services. I hope the standard that emerges isn’t determined merely by the name of the technology involved.

We need to look concretely at whether users can turn off or adjust recommendations, whether they can stop whenever they want, and what default settings apply to minors. Convenience shouldn’t be an excuse for users to lose control over personalization, and the existence of risk shouldn’t mean we treat genuinely useful features the same way we treat harmful ones.

Watching this unfold makes me reconsider what we even use to measure an app’s success. When time-on-app goes up, we need to distinguish whether users are more satisfied, or whether they simply couldn’t finish what they came to do and got stuck longer than intended. If we call both of these “success” using a single number, we risk missing the actual design problem.

I’d like you to think about the apps you use every day. When you try to manage your own screen time, are the tools you need easy to find? Do the limits you set actually help? I think examining that everyday experience is a concrete starting point for talking about what responsibility companies should bear.

References

• European Commission: Preliminary finding on TikTok’s addictive design, February 6, 2026
• European Commission: Investigation into Facebook and Instagram’s protection of minors, May 2024
• European Commission: TikTok commits to permanently withdraw TikTok Lite rewards program, August 2024
• European Commission: Enforcement and fines under the DSA
• Pew Research Center: Teens, Social Media and Technology, 2024
• Nir Eyal: The Web’s Slot Machine — Infinite Scroll and Anticipation of the Next Piece of Content

The author is Oswarld (Kwangseob Ahn). Current roles: Adjunct Professor at Sejong University, Strategy Consultant at INLEVEL9. Career, research, books, and recent work are kept current on the About page. Latest · July 2026: HEMA-2: A Consolidation-Aware Tri-Memory Architecture with Multi-Channel Scheduling for Lifelong Conversational AI.