Europe warns that online platforms expose children to addiction, abuse, sexual content, and manipulation through AI.

The major online platforms and search engines continue to expose children and young people to risks ranging from compulsive use and harmful content to grooming, cyberbullying, AI-generated sexualized images, and access to adult materials, according to the second annual report of the European Board for Digital Services, conducted in cooperation with the European Commission.

In brief

The report covers systemic risks from 23 very large online platforms and two very large search engines. The protection of minors is one of the most recurring issues identified in the report. Risks include platform addiction, sexual or violent content, cyberbullying, grooming, and sextortion. AI increases risks through sexual deepfakes, manipulation of minors' images, and chatbots that can facilitate abuse. The report does not declare the platforms' measures as "best practices" because the DSA is still in the early stages of implementation.

The report analyzes the risks reported by very large platforms, very large search engines, civil society organizations, researchers, trusted flaggers, and extrajudicial dispute resolution bodies. The document covers the period from February 17, 2025, to February 16, 2026, and is based on risk and audit reports published by platforms under the Digital Services Act.

At the center of the report are digital services with over 45 million active users monthly in the European Union. This category includes social networks, online marketplaces, app stores, pornographic platforms, and search engines. The Digital Services Act requires them to identify systemic risks arising from how their services are designed, function, and are used, including from algorithmic systems.

The protection of minors is presented as a major priority. The report notes that, according to a Eurobarometer from 2025, over 9 out of 10 Europeans support actions to protect minors from the negative effects of social networks on mental health, against cyberbullying and online harassment, as well as for mechanisms to limit access to age-inappropriate content.

Risks to children arise from both content and the architecture of platforms. The report mentions compulsive behavior or addiction-like behavior on social networks, compulsive shopping on online marketplaces, exposure to sexual, violent, self-harm, eating disorder content, or unrealistic body standards, as well as participation in dangerous online challenges.

The design of platforms is treated as a risk factor. Features such as infinite scrolling, autoplay, intrusive notifications, highly personalized recommendations, social reward mechanisms through likes, and repetitive short content can extend the time spent on platforms and make it more difficult for young users to disengage. The report notes that minors are more vulnerable due to cognitive and social development, lower digital literacy, and limited ability to assess consequences.

Recommendation systems can amplify the problem. When optimized for engagement and time spent on the platform, they can repeatedly expose minors to borderline but harmful content through accumulation: idealized bodies, extreme fitness, sexualized content, materials about self-harm, or risky challenges. The report emphasizes that sometimes the risk does not come from a single element but from the combination of design, recommendations, notifications, content, and monetization.

Artificial intelligence emerges as a cross-cutting risk. Civil society organizations and some platforms have reported that AI tools integrated into services used by children can facilitate the creation of non-consensual sexualized images, manipulation of minors' photographs, sextortion, cyberbullying, and grooming. The report mentions risks related to sexual deepfakes, "nudification" applications, chatbots, and artificially generated content that can be difficult to distinguish from real content.

The report also describes risks related to child sexual abuse materials. Platforms and civil organizations have reported risks regarding the distribution of CSAM, soliciting or coercing minors to produce such materials, content that sexualizes children or normalizes abuse, as well as sending sexual materials to minors by adults. Private messaging, anonymous or pseudonymous accounts, and redirection to external services are mentioned as factors that can facilitate abuse.

Direct contact between adults and minors is another area of risk. The report mentions grooming, sexual exploitation, solicitation, coercion, sextortion, threats of distributing intimate images, cyberbullying, and harassment from predators or other youths. Some contributions have also drawn attention to parasocial relationships between minors and content creators, in which commercial messages can be perceived by children as authentic personal recommendations.

Pornographic platforms are treated separately due to the risk that minors may circumvent age verification measures and access adult content. The report mentions measures such as age verification, content warnings, blurring of pages, "Restricted To Adults" labels, systems for detecting minors or banned users, as well as automated tools for detecting CSAM materials.

Online marketplaces raise another type of problem: minors may purchase age-inappropriate products or products that should be intended for adults only. The report mentions risks related to adult products, alcohol, tobacco, gambling, drugs, dangerous products, or intoxicants. Some platforms claim to use warning labels, filters, blurring of adult listings, age confirmation pop-ups, and limiting certain interactive activities to adult users.

Risks to minors overlap with other risks identified in the report: misinformation, illegal hate speech, discrimination, gender-based violence, fraud, illegal products, and public health risks. For example, children can be affected not only by sexual or violent content but also by fraud, influencers, advertisements, algorithmic recommendations, and false health information or choices.

The report also enumerates the types of measures used by platforms. These include usage and moderation policies, automated detection and human verification, hash-matching for identifying known child sexual abuse materials, safety settings activated by default for minors, parental controls, limiting messaging functions, restrictions for livestreaming, child-friendly reporting tools, warnings, blurring of sensitive content, and options for managing online time.

For social networks, the report mentions measures such as minimum age requirements, age gates, removal of accounts under the minimum age, management of who can follow minors' accounts, and parental approval for certain settings changes. Some platforms have introduced special accounts for teenagers, higher privacy settings, default deactivation of location services, and limiting public profiles to older users.

However, the report maintains an important caveat: the measures described are not presented as "best practices" or "good practices." The Board and the Commission state that the DSA is still in an early stage of implementation, and the fact that a measure is mentioned by a platform does not automatically mean that it works, that it is applied effectively, or that it ensures compliance with the law.

The document also specifies that it is not a compliance assessment of the platforms and does not replace investigations or enforcement actions by the Commission and national coordinators for digital services. At the same time, the report indicates the direction in which European authorities view the implementation of the DSA: risks are not analyzed only at the level of individual content but at the level of design, algorithms, advertising, monetization, moderation, and business models.

In the perspective section, the report mentions that future editions will benefit from law enforcement decisions and researchers' access to data under the DSA. The document also recalls the first non-compliance decision adopted based on the risk assessment framework of the DSA, against Temu, for failing to diligently identify and assess systemic risks related to illegal products offered on the platform and the harm to consumers in the European Union.

For users, the report shows that online child protection is no longer treated in the EU merely as a parental control or point moderation issue. The risk is related to how platforms are built: what they recommend, what they reward, what they monetize, how long users stay connected, and how quickly they can detect abuse.