Are today’s teenagers “digital natives” or “digital refugees”? While this generation is more tech-savvy than any before, they are also more vulnerable to harm in the digital world. Social media addiction, privacy breaches, and overreliance on AI—when convenience and risk coexist, how many children are being lost in the digital world?

To better understand adolescents’ digital habits and potential risks, the Child Welfare League Foundation (CWLF) conducted a nationwide survey of more than 6,000 junior high and senior high school students[1] . The results, released on August 20 in the 2025 Survey on Children’s Online Safety in Taiwan, reveal that 84.6% of teens use social media daily, with 13.6% already showing signs of addiction. Even more troubling, 8.7% said they or their classmates had shared or leaked sexual images, an increase compared to CWLF’s 2023 survey.

At the same time, the rapid rise of generative AI is reshaping young people’s online behavior. The survey found that 20% of students use generative AI frequently. However, only 39.4% reported verifying the accuracy of AI-generated information, and 40% rarely or never considered avoiding the upload of personal data to AI systems. This suggests that adolescents’ digital privacy remains largely unprotected in the AI era. When facing these risks, only 31.5% had discussed AI safety with parents, and 52% with teachers, highlighting a significant gap in digital literacy education at home and in schools.

As AI opens new possibilities, it also brings unknown dangers for children and youth. Ahead of the new school year, CWLF recommends that the Ministry of Education expand current “Ministry of Education Guidelines for Digital Teaching in Senior High School Level or Below” to include AI-specific lesson plans and ensure their practical application in schools. CWLF also proposes the “5C Actions for AI Education”:
— Care: Pay attention to teenagers’ mental health.
— Connect: Encourage real-life social interaction.
— Check: Verify AI content, sources, and accuracy.
— Control: Review privacy settings and data collection practices when using AI.
— Challenge: Question AI outputs to develop critical thinking instead of accepting all information at face value.

CWLF further urges that the Protection of Children and Youths Welfare and Rights Act be amended to include a dedicated section on “Children’s Rights to Communication and Privacy Protection.” This would establish a specialized online safety authority, define cross-ministerial responsibilities, and require internet companies to conduct risk assessments and reporting. Government and industry must also create rapid response mechanisms to minimize harm to children in digital environments. CWLF calls on all sectors to work together to safeguard the digital safety of the next generation.

[1] The survey was conducted online from May 27 to June 30, 2025, targeting junior high and senior high/vocational school students across Taiwan. Schools were contacted based on the Ministry of Education’s official school list. A total of 6,122 valid responses were collected: 52.1% female, 47.1% male, and 0.8% other; 79.5% junior high, 20.5% senior high/vocational; regionally, 23.6% from northern Taiwan, 24.7% central, 38.5% southern, 11.7% eastern, and 1.5% offshore islands.

Digital Risk 1: 84% of Teens Use Social Media Daily; 13% Show Signs of Addiction

Taiwanese teenagers show extremely high dependence on social media, with 84.6% using it daily. Further analysis reveals that 13.6% already exhibit symptoms of addiction[1], such as constantly thinking about social media, wanting to spend more time on it, neglecting other activities, or using it to escape negative emotions.

A closer look at gender differences shows that the rate among girls (15.5%) is significantly higher than that of boys (11.5%). Compared with international data[2], the prevalence of problematic social media use (PSMU) among Taiwanese teens is similar to the global average. However, Taiwanese boys report higher rates of addiction than their global peers, suggesting that junior and senior high school boys in Taiwan face a relatively more serious issue.

[1] Problematic social media use (PSMU), also translated as “problematic use of social media,” refers to excessive and inappropriate social media use that negatively affects daily functioning (Andreassen et al., 2017; Casale et al., 2023; Shannon et al., 2025; Yang, 2025). This survey adopted the Social Media Disorder Scale (SMDS) from the Health Behaviour in School-aged Children (HBSC) study, with a score ≥5 defined as problematic use.
[2] The HBSC (2023) survey targeted students aged 11, 13, and 15 in more than 40 countries and regions, mostly in Europe and Central Asia. Taiwan’s 13-year-old students reported a 12.8% addiction rate, slightly higher than the international average of 12%. Fifteen-year-olds reported 11.0%, the same as the global average. Taiwanese boys at ages 13 (11.2%) and 15 (11.8%) both exceeded the international average (9% and 8%, respectively).

Digital Crisis 2: Teens Addicted to Social Media Are Five Times More Likely to Leak Sexual Images

According to Ministry of Health and Welfare statistics[1], reports of cases involving the filming or production of child sexual images have sharply increased in recent years, rising from 581 cases in 2017 to more than 3,104 in 2024—a more than fivefold increase in just seven years.

This survey further shows that 6% of students had personally received sexual images of classmates or friends, while 6.8% said their classmates had received such images. Combined, 11.6% of respondents reported that they or their classmates had encountered peers’ sexual images—higher than the 9.7% reported in CWLF’s 2023 survey[2].

In terms of sending such images, 2.2% admitted to sharing their own sexual images with online contacts, and 7% said they knew classmates who had done so. Combined, 8.7% reported that they or their classmates had shared sexual images—up from 5.2% in 2023. CWLF warns that the actual number of victims is likely far higher than official reports, indicating a large “hidden figure.”

Problematic social media use is alarmingly associated with the sharing of sexual images. Further analysis reveals that with each additional grade level, the odds[3] of self-sharing sexual images increase by 19.6%. Male students are also far more vulnerable, with a likelihood of sharing that is 84.6% higher than that of female students. Most strikingly, adolescents struggling with social media addiction face a risk more than five times greater than their peers without such problems.

[1] Statistics from the Department of Protective Services, Ministry of Health and Welfare: https://dep.mohw.gov.tw/dops/lp-1303-105-xCat-cat05.html. Before 2022, records only included filming and producing child sexual content (Article 36, Child and Youth Sexual Exploitation Act). After 2023, distribution and exhibition (Article 38) were added.
[2] CWLF, 2023 Survey on Children’s Online Safety and Digital Media Literacy in Taiwan, conducted via online questionnaires among junior high and senior high students.
[3] Here, “odds” means the ratio of the likelihood of engaging in the behavior to not engaging in it.

Digital Crisis 3: 58% of Teens Use Generative AI, but 60% Rarely Verify Accuracy

Beyond social media, the rapid spread of generative AI introduces new risks. The survey found that 58.4% of Taiwanese adolescents use generative AI at least once a week, and 20.9% use it daily, qualifying as frequent users.

Students primarily use AI for information searches (75.5%), homework support (49.7%), translation (48.2%), summarizing materials (36.0%), and creating images or videos (30.8%). Some, however, admitted to directly copying AI-generated answers to complete schoolwork (12.7%). Many also turn to AI for emotional support—chatting with AI or seeking comfort (28.2%), or asking for meaning in life or advice (16.7%).

Gender differences are notable: girls report higher and more diverse usage (average 4.2 types vs. 3.5 for boys), especially in learning and emotional support. Boys, meanwhile, report higher rates of copying schoolwork (13.9% vs. 11.6%) and coding (10.5% vs. 4.2%).

Safety and media literacy remain weak: only 39.4% of adolescents “often or always” verify the accuracy of AI-generated information, while 40% rarely or never pay attention to avoiding the upload of personal data. Just 13.1% said they often or always rely heavily on AI when solving problems. This indicates that while teens enjoy AI’s convenience, they lack adequate skills in fact-checking and privacy protection.

Parental and school involvement is also limited. Only 31.5% of students reported discussing AI risks with parents, and just 52% with teachers. Further analysis shows that younger students, girls, lower-frequency users, and those less convinced of AI’s reliability are more cautious in their usage. Examining whether adolescents discuss AI risks with parents or teachers provides important insight into their level of caution in using AI. Among these factors, teachers’ reminders are particularly effective, helping to offset the decline in caution seen in students with high trust in AI.

Addressing Teen Social Media Addiction and Preventing Sexual Image Leaks

With digital risks emerging one after another, establishing effective protective mechanisms has become an urgent priority. Research shows that social media addiction not only reflects potential mental health problems but is also closely related to family relationships and parenting styles. This survey further confirms that social media addiction is a major risk factor for the leakage of sexual images among adolescents. CWLF urges parents to observe early signs of excessive dependence on social media, establish clear and age-appropriate usage rules in daily life, and help their children develop healthy online habits.

The survey indicates that adolescents addicted to social media face a significantly higher risk of sharing their own sexual images, with 2.2% of students admitting they had actively sent such images to online contacts. CWLF recommends that the government establish accessible support and counseling services, while also strengthening awareness campaigns on online and social media safety to reduce the risk of teenagers becoming involved in sexual image leaks.

Schools and Parents Must Guide Safe Digital Practices to Protect Children in the AI Era

Although the Ministry of Education’s current “Ministry of Education Guidelines for Digital Teaching in Senior High School Level or Below” includes some AI-related content, the survey shows that only about half of students have ever discussed AI-related risks with teachers. Such discussions, however, significantly improve students’ caution in using AI, particularly in verifying the accuracy of information and avoiding the upload of personal data. CWLF recommends that schools explicitly teach students how to assess the accuracy of AI responses, recognize bias, and apply AI appropriately. It further calls for the Ministry to expand practical lesson plans on AI and ensure they are implemented in classrooms.

Parents also play a crucial role in guiding children through AI-related challenges. CWLF encourages parents to adopt the “5C Actions for AI Education”: Care for adolescents’ mental health; Connect by encouraging real-world social interactions; Check AI content, sources, and accuracy; Control privacy settings and data collection when using AI; and Challenge AI outputs to build critical thinking rather than accepting all information at face value. By putting these five principles into practice, parents can systematically guide their children to develop healthy digital habits—enjoying the convenience of AI while protecting themselves in the online world.

CWLF Calls for Stronger Protection of Communication Rights and Privacy

To protect Taiwanese children and youth from digital harms, CWLF has partnered with civil society groups, scholars, and experts to form a civil coalition advocating amendments to the Protection of Children and Youths Welfare and Rights Act. The coalition calls on the government to adopt UN CRC General Comment No. 25, and to draw on international legislation such as the UK’s Online Safety Act and the EU’s Digital Services Act. It urges the inclusion of a dedicated chapter on “Children’s Rights to Communication and Privacy Protection” in Taiwan’s Act to align with global legislative trends and strengthen online safety mechanisms.

In this future chapter, CWLF and its coalition partners recommend that the following provisions be added:
(1) Establish a central government authority dedicated to online safety for children and youth, with clearly defined inter-ministerial responsibilities.
(2) Require the government to conduct and publish child rights impact assessments on the digital environment.
(3) Mandate internet companies to conduct child risk assessments and reporting.
(4) Require digital products and services to comply with safety-by-design principles.
(5) Create rapid response mechanisms to minimize harm when digital risks occur.
(6) Adopt a “guidance before penalties” principle, developing prevention guidelines before sanctions, with a sunset clause once supporting measures are in place.

◼︎ More CWLF surveys: 
• A Depressed Father's Day? 43% Fathers Struggle, Most Avoid Seeking Help
• Children Score Low on Sexual Knowledge and Need Comprehensive Sexuality Education
• Teens Trapped in Poor Part-Time Working Conditions, One in Four Nearly Scammed