TikTok pulls users into rabbit holes, Brussels could impose fines

In a preliminary finding issued on Friday, November 6, the European Commission stated that TikTok is breaching the EU's Digital Services Act (DSA) because its recommender algorithms pull users into so-called rabbit holes, showing them an ever-growing stream of similar content, including material that can harm their mental health.

According to the investigation, reported by POLITICO, this is the first time the Commission has scrutinized a social media platform's recommender system. Brussels says TikTok failed to properly assess the risk of where the endless stream of personalized videos could lead users, who may be shown sensitive content related to mental health or body image. The mechanism is brutally simple: the more someone engages with a type of content, even harmful content, the more similar videos they receive. According to an EU official, the investigation focused on mental health because that is where the evidence was strongest. The Commission says TikTok's measures so far have been ineffective, and is demanding that the platform pay closer attention to users' explicit signals, such as the "not interested" button, rather than relying solely on how long people watch videos or what they like. "It has to be a real, concrete signal," and not just a minor detail after which the system reverts to its old behavior, one official said of the "not interested" button. Although the DSA requires platforms to offer a non-personalized feed option, only 0.2 percent of users choose it.

This is the fifth time TikTok has come under Brussels' scrutiny, but it has never yet been fined. Under the DSA, penalties can reach six percent of a company's annual global turnover, but the company can review the evidence, defend its systems, and propose remedies — and if the Commission accepts them, it can avoid a fine altogether. That is how the proceeding on advertising transparency ended last December, when TikTok made binding commitments. The Commission had already made a preliminary finding on February 12 of this year that TikTok's addictive design — infinite scrolling, autoplay, push notifications and an overly personalized recommender system — violates EU law. At the time, it also found that screen-time limiting tools and parental controls were ineffective, and that the platform needed to change its basic functioning, for instance by disabling infinite scroll. For Romanian readers, TikTok is a particularly sensitive topic: the Commission opened proceedings against the platform in December 2024, after Romania's Constitutional Court annulled the first round of the presidential election, because Călin Georgescu's campaign had spread extremely quickly on TikTok ahead of the vote, with the help of what appeared to be a coordinated network of accounts.

While TikTok's algorithm pulls users into rabbit holes, Elon Musk's X has turned into a veritable highway for lies, where false claims spread as fast and unimpeded as if that were the platform's actual purpose. According to an investigation published by NewsGuard on October 21, X's community-based fact-checking system failed to flag 93.1 percent of false posts circulating about the U.S. midterm elections, and the same holds true for at least four false claims made by Musk himself, despite the owner having 241.5 million followers. NewsGuard says 55 percent of the false claims it tracked during the election cycle appeared on X, and fourteen of them surfaced there first. According to a 2024 analysis by the Center for Countering Digital Hate (CCDH), Musk published at least 87 posts that year spreading false or misleading claims about the election; these were viewed more than two billion times, and none carried a correction. Brussels has also taken action against X. On December 5, 2025, the Commission fined the platform 120 million euros — the first fine issued under the DSA — because its paid blue checkmark misled users, its advertising database lacked transparency, and researchers were denied access to public data. X has challenged the decision in court. In July of this year, the Commission accepted the platform's remediation plan, although the body representing national regulators called it "overall insufficient." However, the broader investigation opened in December 2023, which also covers the handling of disinformation, has still not been concluded, and this year a new proceeding was launched against X over sexually explicit fake images generated by its artificial intelligence tool, Grok. The cases of TikTok and X both point to the same conclusion: platforms' business models are built on keeping users glued to their screens for as long as possible, and outrage, fear and lies are often more effective tools for that than the truth. So far, EU regulation has tended to warn rather than punish, and each new proceeding shows that platforms only change when the penalty actually hurts.

Ugar