When the interface becomes the problem. Europe starts regulating the design of attention

For years, regulating the internet mostly meant fighting illegal content, protecting personal data, making advertising more transparent, and limiting disinformation. The interface remained almost invisible. We treated it as a neutral wrapper in which a platform delivered content to the user.

That approach is starting to change.

The European Commission has preliminarily found that the way Instagram and Facebook are designed may breach the Digital Services Act. The focus is not a specific post or advertisement, but product mechanisms that keep users engaged. The Commission raised similar concerns earlier in relation to TikTok, pointing to mechanisms such as infinite scroll, autoplay, notifications, and highly personalised recommendations. The context comes from the Commission’s communication on formal action around platform design and its earlier DSA proceedings concerning Meta and the protection of minors.

This matters. The regulator is no longer asking only what a platform shows. It is also beginning to ask how the platform is designed so that we keep looking.

The interface was never neutral

Over the past two decades, many digital products have been developed around a similar set of metrics: active users, frequency of return, session length, retention, and engagement.

From a business point of view, this was logical.

The longer a user stayed on a platform, the more content they consumed. The more content they consumed, the more ads could be shown. The more often they returned, the more data they left behind and the more precisely the experience could be personalised.

In this model, the user’s time stopped being only a measure of interest. It became one of the product’s most important raw materials.

Platforms did not only respond to existing user needs. They began designing environments in which leaving the app required more effort than staying.

This did not happen because of one decision. It was the outcome of thousands of experiments, A/B tests, and optimisations conducted over many years. Each individual change could look harmless. Together, they created a very specific product model.

The attention economy was built into the product

Infinite scroll, autoplay, notifications, and recommendation systems are often presented as features that improve convenience. And each of them can be useful.

The problem begins when they all operate at the same time and are optimised mainly for one goal: reducing the number of moments in which the user might decide they have had enough.

Infinite scroll removes the moment of decision

In the older web, a user reached the end of a page and had to do something: click “next page”, choose another article, or close the site.

That moment was a small friction point, but also a moment of reflection.

Infinite scroll removes it. There is no end of the page. No natural pause. No signal saying: “you have seen enough.” New content appears automatically before the user has time to make a conscious decision.

Technically, it is only a way to load content. Behaviourally, it removes a stopping point.

Autoplay turns choice into the default

Autoplay works in a similar way.

When a video ends, the platform does not ask whether we want another one. It assumes that we do. The absence of a decision becomes consent to continue.

This is a small change in choice architecture. Its consequences are much larger than the appearance of the button suggests.

In a model that requires a click, the user chooses further consumption. In an autoplay model, the user must actively choose to stop it.

Notifications move the product beyond the app screen

We often describe notifications as an information tool. They tell us that someone sent a message, liked a photo, or published something new.

At the same time, they are one of the most important mechanisms for reactivating the user.

The product no longer waits until we open it. It can enter our day, interrupt another activity, and suggest that something has happened which deserves immediate attention.

The more apps compete for the same limited time, the more aggressive this communication becomes.

Recommendations do not only predict interests. They shape them

Recommendation systems were supposed to solve a real problem: there is more content online than any person can browse on their own.

An algorithm helps choose the materials we are likely to find interesting.

But recommendations are not merely a mirror of our preferences. By deciding what we see next, they influence what we become interested in, which behaviours we start to consider normal, and which topics we gradually sink into.

In 2024, the Commission already indicated that Facebook and Instagram algorithms may stimulate behavioural addictions in children and create so-called “rabbit-hole effects”. Current enforcement therefore looks like part of a broader direction, not a reaction to one isolated feature.

The regulator starts looking at the “ingredients” of a digital product

The closest analogy, to me, is the food industry.

A food producer knows that specific proportions of sugar, salt, fat, flavouring, and texture influence how readily we reach for another portion. Each ingredient may be legal and widely used. The problem may appear only at the level of the whole product composition and the way it is consumed.

The digital “ingredient list” of a platform works in a similar way.

Infinite scroll does not have to be harmful in itself. Neither do notifications. Personalisation can save time, and autoplay can be convenient.

But the combination of all these elements, supported by algorithms learning user behaviour in real time, may create a product that is much harder to put down than the assessment of individual features would suggest.

That is why the current regulatory direction matters.

The Digital Services Act imposes additional obligations on the largest platforms, including the assessment and mitigation of systemic risks and the protection of users’ rights and wellbeing. The Commission is directly responsible for enforcing these obligations for the largest platforms operating in the European Union. The rules and listed services are described on the Commission’s DSA page for very large online platforms and search engines.

In practice, this means that a company can no longer look only at whether individual screens or messages comply with the law. It may be asked to explain the effects created by the entire design system.

A platform does not have to intend addiction to design compulsive behaviour

The easiest reaction would be to conclude that platforms consciously decided to addict users.

Reality is more complicated.

Technology companies did not have to write “addict the audience” into their strategy. It was enough to reward teams for increasing retention, session length, return frequency, and interactions for years.

If every experiment is evaluated by whether it increases engagement, the product will gradually evolve towards the solutions that most effectively hold attention.

It does not have to be a conspiracy. It can be a predictable effect of an incentive system.

We see the same pattern in many industries. Organisations optimise for the metrics by which they are managed. Over time, the metric stops merely describing the product and starts shaping it.

That is why responsibility cannot end with declared intent.

The question should not only be: “Did the platform want to trigger compulsive behaviour?” The more important question is: “Could it foresee that effect, did it measure it, and did it take meaningful steps to limit it?”

Would regulating the interface limit innovation?

The strongest counterargument is obvious: the features named by regulators also create real value for users.

Infinite scroll can make browsing easier. Autoplay can make watching content smoother. Recommendations help people find valuable materials in a huge database. Notifications inform us about messages and other people’s activity.

Banning all of these functions would be too simple.

Regulation does not have to mean every platform returns to a “next page” button. The more likely shift is towards proportionality, user control, and a real ability to end a session.

Are engagement mechanisms on by default? Can the user turn them off easily? Does the platform analyse the risk of excessive use? Does it design experiences differently for children? Does the recommendation system maximise only the probability of the next interaction, or does it also consider quality and safety?

These questions are harder than simple bans. They are also closer to how digital products are actually built.

From time spent to quality of time spent

For years, high time spent in an app was presented almost automatically as proof of success.

It is not always that.

A longer session can mean the product delivers real value. It can also mean that the interface makes it difficult to stop, recommendations lead to more content, and the user closes the app without remembering why they opened it in the first place.

Time alone does not tell us which interpretation is true.

The industry may therefore need new measures.

Not only: how many minutes did the user spend in the product?

Also: did they achieve their goal, did they feel in control, did they return because they wanted to or because they were effectively reactivated, and how did they feel after the session ended?

These metrics are less convenient than time, page views, or retention. They are also harder to connect to quarterly financial results.

But they may describe product quality better.

A new question for product designers

The most important consequence of the Commission’s actions will probably not be the banning of one feature. It will be a shift in how responsibility for design is understood.

Until now, a designer, product manager, or growth manager could assume that if a mechanism was legal, effective, and increased engagement, it could be shipped.

In the future, one more question may be necessary:

At what cost do we achieve this growth?

This question does not apply only to the largest social platforms.

It applies to news sites, shopping apps, streaming platforms, games, education tools, and almost every product competing for limited user attention.

For years, the industry’s basic question was:

How do we keep the user with us for five more minutes?

Increasingly, the next question may be different:

How much of the user’s attention are we allowed to take — and how must we justify the way we do it?

The interface is no longer an innocent wrapper around technology. It becomes part of the product, the business model, and the company’s responsibility.

And time spent in an app may soon stop being only a point of pride. It may also become a question that companies need to answer to users, boards, and regulators.