Evidence Library
Comprehensive data charts, academic papers, and statistical analysis on the impact of social media on children and adolescents.
Percentage of websites and apps with unsuitable design features by type of design (2025)
This chart illustrates the prevalence of high-risk design features among 876 digital services analysed in the 2025 GPEN Sweep. The data reveal that 20% of platforms allow children to freely engage with others without adequate friction or oversight, while 15% subject minor users to behavioural profiling. The methodology also reveals a significant barrier to transparency: 15% of services use privacy policies written in language that is likely beyond the comprehension of child users. A notable trend emerges concerning the relationship between business models and child safety. The Sweep found that mobile apps are notably riskier than websites, with an unsuitability rate of 47% compared to 41%. Furthermore, a 'monetisation gap' was identified: free services, which often rely on data-driven advertising, were found to be suitable in only 53% of cases, whereas paid/subscription services showed a higher suitability rate of 61%. Free of fees platforms create incentives for the collection of intrusive data (e.g. 11% geolocation and 12% public-by-default settings) and the use of engagement features.
Perception of efficacy of a social media ban in six European countries (2026)
This chart illustrates public perceptions of the efficacy of banning under-16s from social media across six European countries. The data reveal an efficiency gap: while a majority of Europeans support such a ban in principle, they are deeply divided and often pessimistic regarding its ability to prevent children from creating and using accounts. A significant regional trend emerges in terms of optimism levels: Italy is the most confident about the efficiency of the policy, with 61% of adults (25% of whom believe it will be "very effective") expecting the measure to succeed. Conversely, Britain is the most sceptical, with 54% of adults in total predicting that the ban will fail, including 17% who believe it will be 'not at all effective'. This scepticism is also evident in Poland and Germany, where approximately half of the population doubts the practical impact of the law.
Perceptions of the importance of digital skills
This table illustrates how the public perceives the importance of digital skills in relation to social, economic and educational factors in the European Union. The data reveal a near-universal consensus that digital literacy is a fundamental utility, with 89% of respondents ("strongly" and "somewhat") agreeing that these skills are essential for basic participation in society, for example, in banking and healthcare. The importance of digital competence is also recognised in the areas of education, with 85% of respondents ("strongly" and "somewhat") agreeing that these skills are essential for navigating the modern economy. The data reveal that 40% of respondents strongly agree that digital skills are essential for formal education and for the safe and responsible use of generative AI tools such as ChatGPT and Gemini. Furthermore, 80% of the public identifies digital literacy as a prerequisite for a successful career and a defence against misinformation.
Perceptions of the the impact of digital technologies
This table shows the European population’s perception of the impact of digital technologies on education and training. The data reveal a broad consensus on the benefits of digitalisation, particularly regarding innovation and personalisation. A combined 87% of Europeans agree that technology makes education more innovative (35% strongly agree and 52% somewhat agree), and 77% agree that it allows for tailored learning (33% strongly agree and 54% somewhat agree). The survey also highlights a strong belief in the capacity of digital tools to foster inclusion and accessibility. A total of 84% of respondents agree that these technologies can break down barriers for people with disabilities, with 36% strongly agreeing: the highest proportion in the dataset. Similarly, 83% agree that technology makes education more accessible and inclusive overall (31% strongly agree and 52% somewhat agree), and 85% see it as beneficial for lifelong learning (33% strongly agree and 52% somewhat agree). However, the data also reflects significant socio-economic concerns. While the level of agreement is lower than for the benefits, a substantial 67% of Europeans still agree that digital technologies are too expensive for families (21% "strongly agree" and 46% "somewhat agree"). Furthermore, 64% of respondents express concern that these tools can increase inequalities in education (21% "strongly agree" and 43% "somewhat agree").
Preferred information sources for social and political current affairs
This table shows the percentage of people who prefer to get their information about social and political current affairs from the sources listed. TV remains the dominant medium, with 71% of respondents identifying it as a primary source. A significant secondary cluster of sources is closely followed, with nearly equal preference levels: radio (43%), the printed or online press (41%), search engines (40%), social media platforms (40%) and personal networks such as friends and family (40%). Digital-specific and emerging formats are less popular overall, with video platforms cited by 26% of respondents, followed by podcasts (15%) and AI chatbots (9%). Formal institutional sources, such as school teachers or university lecturers, are among the least popular options, with a mere 6% preference. The data indicates a clear hierarchy in which traditional broadcast media still lead, while a broad range of digital and interpersonal sources have consolidated around a 40% share of public preference.
Prevalence of Age Assurance Mechanisms Among Online Services (2024)
Based on 50 digital service instances. Please note that services may employ more than one mechanism, so the totals represent the frequency of use across the sample. This chart illustrates the prevalence and systematic nature of age assurance mechanisms across 50 online services in 2024. The data represent the percentage usage within each category. The data reveal that 'age verification in some cases' (conditional/situational checks) is the most widely used technical strategy. This approach is dominant in the pornography sector (80%) and social media (50%), suggesting that these high-risk areas rely heavily on reactive verification triggers. A significant trend is the continued reliance on the 'honesty box' model: 'Self-declaration only', in fact, remains a primary tool for several sectors, accounting for 66.7% of immersive environments and 50% of both generative AI and 'for kids' services. Additionally, 'age estimation in some cases' serves as a frequent situational layer, particularly for social media (42.9%) and random live video chat (28.6%). The data also highlight the extreme rarity of systematic barriers across the entire user base. 'Age estimation in all cases' is used by only two sectors: Dating (25%) and Gaming (10%). Furthermore, the chart reveals a significant enforcement gap in the App Store category, which has the highest rate of total non-compliance, with 40% of services providing neither age verification mechanisms nor self-declaration prompts. Overall, these findings emphasise that technical interventions are currently used as discretionary filters rather than as universal entry requirements, even in services designed specifically for children or containing high-risk content.
Proportion of 13–17-years-old who have seen different types of violent content on social media in the past 12 months
This table shows the prevalence of specific violent and illegal content that 13- to 17-year-olds have encountered online. A total of 70% of children report having seen some form of violence, with fights involving young people (56%) and threats of physical harm (43%) being the most commonly reported experiences. Significant exposure to illegal or high-risk activities was also reported, with 38% of young people seeing content related to illegal drug use and 35% seeing the carrying or use of weapons. Furthermore, 33% of respondents report seeing the promotion of gangs or illegal drugs. Even the least prevalent categories, such as glorifying previous attacks (29%) and sexually violent content (27%), affect over a quarter of those surveyed, highlighting widespread exposure to various digital risks.
Proportion of children who have perpetrated violence who say that social media has played a role
This table illustrates the influence of social media mechanisms on adolescent violence. While 35% of respondents report no connection, a significant 64% of perpetrators cite digital triggers as the main cause of their behaviour. The data reveal that the most common catalysts for physical aggression are online arguments (23%) and 'online disinhibition,' whereby children feel safer saying things they would not say face-to-face (23%). Furthermore, features such as tagging peers into conflicts and live-streaming fights demonstrate how platform design can be used to broadcast and intensify violence, transforming digital disputes into real-world harm.
Public Authorities Priorities Related to Digital Technologies
Almost nine out of ten European respondents (88%) state that countering and mitigating the issue of fake news and misinformation online. 83% think it is important for public authorities to shape the development of Artificial Intelligence and other digital technologies to ensure they respect our rights and values.
Public Opinion on Banning AI Tools Use in Schools (2024)
The data reveals a clear "hardening" of attitudes toward AI in the classroom. Within a single year, support for an outright ban on AI tools in schools has increased from 29% to 36%. Conversely, the proportion of the population that once opposed a ban has dropped to 37%, leaving the global public almost evenly split. A significant 27% of the population remains undecided. The survey covers 30 countries: Argentina, Australia, Belgium, Brazil, Canada, Chile, Colombia, France, Germany, Great Britain, Hungary, India, Indonesia, Ireland, Italy, Japan, Malaysia, Mexico, the Netherlands, Peru, Poland, Romania, Singapore, South Africa, South Korea, Spain, Sweden, Thailand, Türkiye, and the United States.