TikTok faces EU finding over For You feed rabbit holes | Magica
EU says TikTok’s For You safeguards fail to stop harmful ‘rabbit holes’
Editorial Team
••📖5 min read
EU regulators preliminarily found TikTok failed to adequately assess and reduce risks from repetitive video recommendations. The case challenges engagement signals and ineffective user controls, including for legal content. TikTok disputes the findings.
EU regulators say TikTok inadequately assessed and reduced risks from repetitive recommendations.
The case examines how legal videos accumulate in a feed, extending beyond content removal.
TikTok can challenge the evidence before a final decision; confirmed violations could bring a revenue-based fine.
The European Commission preliminarily found on November 6 that TikTok breached the Digital Services Act over harmful content loops in its For You recommendation feed. The EU’s executive arm says the short-video platform failed to adequately assess and reduce the risks of repeated recommendations, particularly for vulnerable adults and minors, according to its announcement.
The stakes reach into the ranking system of the ByteDance-owned platform: regulators want recommendations to give greater weight to what users explicitly request, rather than infer their preferences from viewing time. TikTok disputes the findings, which are not a final ruling.
Officials said investigators analysed around 20,000 videos. Confirmed violations could bring a fine of up to 6% of annual global revenue; neither figure measures users harmed. Source: The Record.
The risk lies in repetition
For You supplies a continuous stream of personalised videos. The Commission says interaction with a subject brings more similar recommendations, potentially exposing vulnerable adults and minors repeatedly to material about sadness, isolation or body image.
Officials deliberately examined videos that were legal and permitted under TikTok’s own rules, according to a report on their briefing. They cited trends such as “what I eat in a day” and “SkinnyTok” to explain how individual clips could appear harmless while repeated exposure could become damaging. The distinction matters for enforcement: deleting prohibited posts would not by itself address the behaviour under examination.
Officials said TikTok recognised some concerns in its own risk assessments but did not adequately analyse them or establish effective safeguards. They said investigators obtained internal documents and data and analysed around 20,000 videos with help from psychologists, psychiatrists and system-design experts. They also said they had enough information without access to the exact algorithm.
Chip shares recovered part of Thursday’s losses in early Friday trading. Fears about data-centre construction face a test against AI-chip supplier Nvidia’s strong reported sales, growth forecast and conditional infrastructure-financing plans.
That number describes material examined, not the proportion of TikTok users experiencing harmful loops. Officials withheld specific testing methods to prevent platforms from gaming their tools, leaving an important limit on outside scrutiny of the findings.
TikTok’s spokesperson rejected the Commission’s account in a statement carried in the same report, saying it misrepresented the platform’s architecture and safety efforts. The company also defended its record of protecting users and providing direct, intuitive control over their content experience.
Existing controls face an effectiveness test
TikTok has already made changes under the DSA, the EU rulebook that requires platforms to address online risks and gives regulators powers to impose substantial penalties. In an August 2023 compliance announcement, TikTok said European users could turn off personalisation in For You and LIVE feeds and receive locally relevant and globally popular videos instead. It also described a refresh tool and chronological Following and Friends feeds.
The present dispute is whether those protections work adequately. Officials said techniques intended to diversify recommendations were ineffective, refreshing a feed helped only briefly, and the “not interested” control lost effectiveness over time.
Regulators want more accessible controls and changes to the signals that determine recommendations, including ending reliance on viewing time as a recommendation metric, according to the account of their findings. That would move the remedy beyond adding another button to changing how a personalised feed interprets behaviour.
The potential reach is substantial. TikTok said its EU community numbered 169 million people in an August 2025 transparency announcement accompanying its report for the first half of that year. That company-provided figure is a historical measure of scale, not an estimate of users harmed.
The same announcement said TikTok removed around 27.8 million pieces of content violating its Community Guidelines during the first half of 2025. That figure measures removals, not whether repeated recommendations of permitted content are safe—the distinction at the centre of this case.
TikTok was designated a very large online platform under the DSA on April 25, 2023, following its declaration of 135.9 million monthly active EU users. That status brought obligations to assess and mitigate systemic risks arising from its design, explaining why its recommendation system faces this scrutiny.
They follow separate preliminary findings this year concerning addictive design in February and protections for minors’ accounts in July. Officials distinguished the latest case from issues such as notifications and endless scrolling: it concerns the recommendation system itself.
Officials described this as the EU’s first direct challenge to a major social platform’s engagement-based recommendation system. They also acknowledged that other platforms produce similar loops and said the Commission was looking beyond TikTok. The claimed enforcement novelty therefore does not establish that the underlying design problem is unique to this company.
TikTok’s response precedes a final decision
TikTok can examine and challenge the Commission’s evidence before regulators decide whether to confirm the violations. A confirmed breach could bring a fine of up to 6% of annual global revenue. The DSA sets no legal deadline for concluding formal proceedings, as the Commission explained when opening the investigation.
There is also a wider policy stake. Officials said the investigation inspired a provision in the Commission’s proposed KIDS Act, introduced in September, that would prohibit platforms from relying solely on engagement signals to recommend content to children, according to the briefing account. That is a proposed rule, distinct from the current DSA enforcement case.
The central decision is whether the Commission can substantiate that TikTok’s risk assessment and safeguards fall short—and what changes to ranking and user controls would adequately address that failure. A preliminary finding sets out the regulator’s case; TikTok’s response and the final decision will determine its outcome.
SoftBank shares fell 10.8% as default-insurance costs reached a four-year high. Delayed listings and unfinished data centers complicate its AI investment outlook, while a $40 billion bridge facility has a March 2027 maturity and a repayment plan based on existing assets and other financing.
U.S. prosecutors allege a global fraud combined AI impersonation, sham confidential deals and cryptocurrency laundering. A blocked bank transfer shows why convincing an employee and moving the money are separate hurdles.
President Donald Trump pledged support for states embracing AI without announcing project commitments. Existing federal incentives and Virginia’s data center review show why the returns depend on tax concessions, construction and who pays for electricity infrastructure.
Tesla plans to wind down Cybertruck production at its Texas factory around mid-2027 to expand Cybercab output. Weak pickup sales frame the retreat, while battery constraints and rising investment complicate the robotaxi expansion. Final truck output and customer order handling remain unclear.
Activists breached the Amsterdam construction site intended for Microsoft, joining opposition to data center expansion with allegations about surveillance. The developer plans more than €1 billion in investment, while government records acknowledge competition for scarce grid capacity.
Airbnb plans to start rolling out an AI travel agent in the first half of 2027. Its wider ambition to bring partner bookings into its app could bring more revenue and support obligations, while Expedia’s abandoned Romie assistant shows the execution risk.
South Korea halted KOSPI share trading after a plunge of more than 8%, while Japan’s Nikkei touched 20% below its June peak. Investors fear a slowdown in AI infrastructure spending, but the market losses do not yet establish cuts to construction or supplier orders.
Kalepa’s $75 million Series B will fund more insurance workflows, a planned London office and a doubling of staff in 2027. The company says it has been cash flow positive for over two years; its next test is extending automation while preserving insurer control over consequential decisions.
The Cambridge startup has secured backing described as eight figures in capital and compute, including UK government support. Its wager is that cheaper, accurate quantum calculations can supply better AI training data; full performance results remain forthcoming.
Japan’s Nikkei fell 7.09% in Friday morning trading, with AI, chip and data center-related shares hit hard after Wall Street losses. Reports linked the rout to fears of U.S. regulation following Democratic election victories, without identifying an enacted restriction or a cut in AI spending.
Potential investors are pressing Anthropic for a lower valuation as a technology selloff threatens its November IPO timetable. Rapid revenue growth and substantial cash complicate a financing debate driven by losses and long-term infrastructure commitments.
Visa's new benchmark finds that AI shopping models often miss customer requirements and struggle to repeat successful decisions. The tests cover product discovery, leaving checkout and payment reliability for future rounds.
Agility Robotics began trading after its Churchill Capital Corp XI merger, bringing a commercially deployed humanoid robot maker to U.S. public markets. Its next test is turning conditional Digit 5 orders into deliveries and sustainable economics.
Goldman Sachs sees temporary halts potentially affecting about 45% of planned US data center expansion after Democratic election victories. In Texas, where approvals were already paused, the next fight is over how long restrictions last and what developers must pay for.
Core Scientific’s third-quarter colocation revenue rose 57.6% to $215.4 million. Its expansion beyond CoreWeave could diversify sales from 2027, but the next projects require additional financing rather than the same customer funding that supported its existing conversions.
Meta is rolling out voice conversations and animated avatars for Muse in the US. The update adds personality to a cloud-based task agent whose costs, permissions and reliability matter more than its expressions.
Anthropic is rolling out Personal Finance to paid Claude users in the US, using Plaid to read bank data. Its pitch rests on ad-free answers and privacy controls, while its own documentation warns that balances and budgets can be incomplete.
OpenAI president Greg Brockman predicts significant medical advances within 12 to 18 months and raises the possibility of death becoming a choice. OpenAI’s cell research and a separate early drug trial show narrower progress, with financial ties, safety questions and clinical testing still central to judging the promise.
Citrini Research is recasting its AI investments around relative value as it warns that financing and construction constraints threaten returns. Its reported Oracle–Nebius trade faces a crucial counterargument: Nebius has substantial customer commitments, although delivery remains conditional.
