Opening Statement and Submission to the Joint Oireachtas Committee on Children and Equality

Screen time often receives a lot of attention. However, Irish and Ireland-linked evidence indicates that “screen time” on its own is an imprecise predictor of harm. Associations with wellbeing tend to be small and inconsistent when measured robustly, while poorer outcomes are more consistently linked to specific online experiences and mechanisms, particularly cyberbullying/peer harm, sexting behaviours (including pressured or unwanted requests and non-consensual sharing), appearance related distress, and sleep displacement or sleep difficulty (Orben & Przybylski, 2019; Foody et al., 2021; Foody et al., 2023; Cotter et al., 2025; Walsh & Perrotta, 2019).

While Irish population-based studies of adolescents have found high social media use, body dissatisfaction, and sexting behaviours were each associated with poorer mental health and elevated self-harm risk indicators, often stronger for girls (Cotter et al., 2025), the authors refrain from claiming these associations as causal.

Although debates are often framed in sometimes medicalised language such as dosedependent responses, “dopamine hits” , and addiction, the scientific evidence suggests a far more complex picture for young people (Meshi and Binder, 2025). This involves a complicated interrelationship between individual characteristics, school and family contexts, in addition to broader cultural and sociological factors (Etchells, 2024). Best practice randomised controlled trials and observational studies have been used to evaluate the impact of smartphone bans and digital detoxes across recent studies. The overall findings are quite mixed, but a large portion of research for school bans showed no significant changes in the mental health and well-being of participants, their average time spent on social media or their educational outcomes.

Online risks change as children grow, but peer harm, particularly bullying and cyberbullying, remains a consistent concern across age groups.

Due to different methodologies and definitions, estimates of bullying in Ireland can vary widely, making it hard to build a clear overall picture. For example, a recent nationally representative survey of 2nd-year post-primary students found that 6.4 % reported online bullying victimisation at least once a month, with boys (9.6 %) more than twice as likely as girls (3.8 %) to report monthly online victimisation (O’Higgins Norman et al., 2026, in review). In contrast, the Health Behaviour in School-aged Children (HBSC) Ireland 2022 Study, published in 2024, reports that about 15-18 % of Irish adolescents aged roughly 10–17 have experienced cyberbullying (being cyberbullied) at least once in the past couple of months, reflecting a broader definition and older age range.

In a separate bystander study in Ireland, roughly 60 % of young people reported having witnessed cyberbullying on social networks, highlighting that many encounter online bullying even if they are not direct targets (Sanmartín et al. 2023).

Either way, these data underscore that online bullying is a significant minority experience among young people in Ireland, and that bystander exposure, where young people witness these behaviours, is even more common, making bystander dynamics a critical component of understanding bullying in the digital age.

Other recent research by DCU Anti-Bullying Centre highlights the growing impact of toxic and misogynistic online influencers, particularly within short-form video platforms such as TikTok and YouTube Shorts, where algorithmic recommender systems play a central role in amplifying harmful content. Experimental research using simulated adolescent user accounts found that boys were exposed to male-supremacist, anti-feminist and reactionary content within minutes of use, even when accounts initially engaged only with generic interests such as sport, fitness or gaming. Once engaged, algorithmic systems rapidly increased the volume and intensity of this content, with the majority of recommended videos becoming toxic within a short viewing period. This content frequently exploits boys’ vulnerabilities around mental health help-seeking, identity and economic insecurity, promoting rigid and oppressive models of masculinity that normalise misogyny, emotional suppression, aggression and entitlement.

These dynamics intersect with emerging and deeply concerning risks linked to image-based sexual abuse and AI-generated child sexual abuse material (CSAM), including the recent proliferation of so-called “nudification” tools that enable the creation and circulation of non-consensual sexually explicit images, and particularly of girls and minors (Baker et al, 2024; O’Rourke et al, 2024).

Elements of both misogynistic influencers and so-called “nudification” technologies represent a intensified form of sexual violence which are underpinned by entitlement-based conspiratorial ideologies often promoted within segments of wider society. Evidence from experimental research, schools and youth settings indicates that these narratives are shaping how some boys perceive and interact with girls and women, contributing to the normalisation of sexist language, sexual harassment and gender-based abuse, with serious implications for girls’ safety, dignity and participation in educational spaces (Renström and Bäck, 2024; Wescott et al., 2024).

Taken together, the findings underline that algorithmically amplified toxic influencer cultures present a dual harm: they undermine boys’ emotional wellbeing, empathy and healthy identity development; while simultaneously escalating the risk of sexualised and gendered harms to girls, reinforcing the urgent need for robust regulation, safeguarding-by-design, and accountability for platforms and AI systems that enable and amplify these abuses (Flynn et al., 2025; McCashin, 2024; Setty et al., 2025). In Ireland, there is also a strong appetite from the education sector, youth justice sector and parents groups to engage with evidenceinformed toolkits that facilitate meaningful dialogue with young people on such issues, not limited to: digital citizenship, masculinity and identity, sex and sexuality, and combatting mis/disinformation.

It is very important to note that frontline practitioners in cybercrime and researchers alike have long anticipated the misuse of genAI for the offences we are witnessing in recent times. The high-profile coverage of Grok has raised extremely serious concerns for everyone in this room regarding the online safety and dignity of young people online. Since September 2025, The Observatory project within DCU Anti-Bullying Centre, funded by the Department of Justice, has been investigating the role of emerging technologies in child sexual abuse material (CSAM) and will be publishing an evidence review later this year. Although empirical research specifically for Grok is limited, we would like to highlight some research insights which may be of note.

A recent report by AI Forensics collected 50,000 mentions of Grok and 20,000 images generated by Grok between December 25th and January 1st (inclusive). It found: 53% of images generated by Grok contained individuals in minimal attire of which 81% were women; 2% of images depicted persons appearing to be 18 years old or younger (as determined by Google's Gemini vision model); and 6% of images depicted public figures, approximately one-third of whom were political figures. Overall, a range of approximately 74-92% of the images were of females, generated by profiles that were 83% male-presenting in terms of profile name/picture. Updated data from AI Forensics as of January 13th and 14th suggest that these figures have significantly reduced in response to recently introduced safeguards. However, Grok can still be used via the website and app to generate far more explicit images than on X (AI Forensics, 2026). Unfortunately, there is also a wider range of similar technologies being abused for nefarious purposes by those with specialist skills (Easttom, 2025).

Therefore, it is reasonable to suggest that there is likely a range of typologies of individuals using such technologies for differing and potentially overlapping reasons. This includes, but is not limited to: those motivated by commercial interests, the (gendered) exertion of power and control, fantasy and/or contact-driven sexual interests, or those who purport to be ‘curious’ in exploring the technology based on what is a clear misperception of the harm caused by such AI (Fido, 2026; Flynn et al., 2025; Harper et al., 2021; McGlynn and Toparlak, 2025). It is important to note that pre-teen children who are developmentally incapable of implicitly understanding the risks of such technologies are at risk of becoming perpetrators of real harm to others, especially if these technologies are being advertised to them.

Of the factors associated with the perpetration of image-based sexual abuse, evidence reviews have found that the acceptance of myths that minimise/excuse the perpetrator correlated positively with the perpetration of non-consensual distribution of intimate material and all forms of image-based sexual abuse (Paradiso et al., 2024). Therefore, it is crucial that future research understands the factors that influence how young people either accept or reject the many myths associated with image-based sexual abuse. Finally, scholars have continually noted the absence of impactful victim-centric responses to image-based abuse, such that many of those targeted by this imagery have reported feeling unsupported (Caletti and Summerer, 2024).

Research consistently identifies a persistent tendency among children and young people not to report bullying, alongside particular weaknesses in students’ self-efficacy to report harm when they witness unwanted behaviour online. Studies indicate that a substantial proportion of adolescents act as passive online bystanders, with between 45.3% and 58.8% reporting having witnessed cyberbullying without intervening, highlighting that many young people encounter online bullying indirectly rather than as direct targets. Evidence further shows that online bystanders are often less confident in recognising, interpreting, and knowing how to respond appropriately when cyberbullying occurs, reinforcing non-reporting as a significant barrier to effective intervention (O’Higgins Norman et al., in review; O’Higgins Norman et al., 2024; Feijóo et al., 2025).

Our research indicates that neither primary nor post-primary students have been meaningfully consulted on proposed smartphone bans and other issues related to social media and the internet, nor has their feedback been systematically sought in contexts where such bans have already been implemented. As a result, children and adolescents are once again excluded from decisions that directly affect their daily lives. This raises significant concerns in light of the rights afforded to children and young people under Article 12 of the United Nations Convention on the Rights of the Child, which affirms their right to be heard in matters affecting them (Reynolds, et al. 2025).

Similarly, the framing of a ‘social media ban’ belies the reality that any proposed ‘ban’ is in fact a requirement for industry to meet safety thresholds to ensure it is age-appropriate for young people. Current bans in other countries (such as Australia) do not include gaming and gen-AI platforms, and have practical challenges in effective implementation. In keeping with many of the issues raised here, there is very poor evidence that social media bans work (Livingstone, 2026; Radtke et al., 2022), and recent real world evidence from UK schools with over 1,200 students found no evidence that restrictive school policies were associated with overall phone and social media use or better mental wellbeing in adolescents (Goodyear et al., 2025).