Growing up in the online world: a national consultation “We are consulting on further measures to prepare children for the future in an age of rapid technological change. This includes potential age restrictions on social media and other services such as gaming sites and AI chatbots, restrictions on addictive design features and risky functionalities, and better support for parents and families.” UK GOVT

 Introduction

Social media companies increasingly rely on highly personalised algorithms to maximise user engagement and advertising revenue.[i] By analysing behavioural data such as viewing time, clicks, and interactions, AI-driven recommendation systems predict and prioritise content most likely to retain user attention.[ii] As more user data is generated, these systems have become increasingly sophisticated, prompting growing concern regarding their influence on mental health and wellbeing. In March 2026, a landmark trial concluded that Meta and Google had intentionally designed addictive platform features that contributed to harm to a young adult’s mental health, with particular concern raised regarding the impact of algorithms on children and young people.[iii] This is especially pertinent in the UK, where Ofcom research found that eight in ten children aged 8–17 use at least one social media platform, while over half of children aged 3–12 reportedly use social media applications despite minimum age restrictions.[iv]

Unlike earlier forms of social media that primarily displayed content from accounts users actively followed, modern platforms increasingly curate feeds through AI-driven recommendation systems. Rather than relying solely on user choice, algorithms continuously shape content exposure based on engagement behaviours, often introducing material from outside existing social networks.2 Research has demonstrated that simulated accounts posing as young teenagers can begin receiving recommendations relating to self-harm, eating disorders, or weight-loss content within minutes of engaging with similar material.[v] Consequently, there is increasing recognition that the risks associated with social media are not simply linked to platform use itself, but to the algorithmic systems that curate, amplify, and repeatedly reinforce certain forms of content. While social media companies argue that personalisation improves user experience and online connection, critics contend that engagement-driven systems may intensify exposure to harmful material and contribute to adverse mental health outcomes among children and young people.[vi] Despite growing public and political scrutiny, regulatory responses have struggled to keep pace with the rapid development of AI-driven technologies. In the UK, legislation such as the Online Safety Act represents a move towards greater platform accountability, although questions remain regarding its effectiveness in addressing the specific risks posed by algorithmic content curation.[vii]

The Rise and Evolution of Algorithms

The use of algorithmic curation on social media is not new; however, its scale, sophistication, and influence have evolved significantly over time. Early social media platforms such as Facebook initially presented content in largely chronological order, where users viewed posts exclusively from accounts they chose to follow.[viii] While this model was not without harm, with issues such as cyberbullying and exposure to unrealistic standards were being prevalent, it nonetheless allowed users a degree of control over the content they encountered.

As platforms expanded and user bases grew, chronological feeds became increasingly saturated, prompting a shift towards algorithmically ranked content. By the mid-2010s, platforms began prioritising posts based on predicted relevance rather than recency, using behavioural signals such as likes, shares, and time spent viewing content.6

The introduction and rapid advancement of artificial intelligence further transformed these systems. Machine learning models enabled platforms not only to rank content from existing networks, but to actively recommend new material beyond a user’s immediate connections.[ix] This shift is most clearly exemplified by platforms such as TikTok, where the “For You” feed is almost entirely driven by AI-powered recommendations rather than follower networks. Competing platforms, including Instagram and YouTube, have since adopted similar models in response to changing user expectations and market pressures.[x]

From a commercial perspective, this evolution is highly rational. AI-driven recommendation systems maximise user engagement by continuously refining content based on real-time behavioural data, thereby increasing time spent on platforms and exposure to targeted advertising, with 88% of marketers engaging with AI to achieve higher customer engagement across platforms.[xi] In a competitive digital economy, where user attention is a primary commodity, these systems offer a powerful mechanism for growth and monetisation.

However, this shift also represents a fundamental risk to the safety and wellbeing of users, particularly children and young people who are immersed in a continuous stream of generated, personalised content which they have limited visibility or control over.

Impact on Children & Young People

The increasing integration of AI-driven algorithms into social media platforms has been associated with a range of negative mental health outcomes among children and young people. While social media itself has long been linked to issues such as anxiety, depression, and low self-esteem, algorithmic curation intensifies these risks by shaping not only the quantity, but the nature and repetition of content that users are exposed to.[xii]

One of the most significant concerns is the amplification of harmful content. Research has demonstrated that engagement-based algorithms tend to prioritise material that elicits strong emotional responses, including content related to body image, self-harm, and disordered eating.[xiii]  This was reflected in the inquest into Molly Russell, a 14-year-old who was exposed to large volumes of self-harm and suicide-related content on Instagram and Pinterest prior to her death. The coroner concluded that the algorithmic curation of this material contributed to her mental health deterioration.[xiv][xv]

This process is underpinned by the feedback loop inherent in AI-driven systems. As young people engage with content, whether intentionally or passively, the algorithm interprets this behaviour as preference, further refining and narrowing the content presented.2 Over time, this can create an “echo chamber” effect, where users are repeatedly exposed to similar themes or viewpoints, limiting exposure to diverse or protective content. For children and adolescents, whose identities, beliefs, and self-concept are still developing, such environments may have a disproportionate influence on emotional wellbeing and social comparison.[xvi]

Beyond content exposure, algorithmic systems are designed to maximise user attention through features such as personalised feeds, infinite scrolling, and autoplay. [xvii] These features encourage prolonged engagement and habitual use, which have been associated with disrupted sleep, reduced physical activity, and poorer mental wellbeing.[xviii] Approximately 64% of 8–14-year-olds report using devices between 11pm and 5am at least once within a four-week period,[xix]  while research has found that device use within 90 minutes of sleep is associated with significantly poorer sleep outcomes.[xx] As sleep is essential for emotional regulation, cognitive development, and physical health, the continuous nature of AI-driven feeds may contribute to a cycle of compulsive use that further exacerbates mental health risks among young people.[xxi][xxii]

Although children and young people are often considered particularly vulnerable to algorithmic harms due to ongoing cognitive, emotional, and social development, many of the mental health implications associated with AI-driven social media algorithms extend to adults. Research has linked algorithmically curated content to increased anxiety, depression, loneliness, body dissatisfaction, and problematic social media use across age groups.2 However, evidence suggests that these impacts are not experienced equally. Young women appear disproportionately affected by appearance-focused content and body image pressures,[xxiii] while young men may be more vulnerable to algorithmic exposure to misogynistic, extremist, or “manosphere” content.[xxiv] Social and economic factors may also influence risk, with individuals experiencing social isolation or limited access to offline support networks more likely to rely on online communities and AI-mediated forms of support.[xxv] While these technologies may provide new avenues for connection and help alleviate feelings of loneliness, they do not replace the importance of meaningful human relationships, which remain fundamental to psychological wellbeing and long-term health.[xxvi] Consequently, algorithmic systems have the potential to reinforce existing social and health inequalities, meaning that the benefits and harms of AI-driven social media may vary considerably across demographic groups.

However, the mental health implications of algorithmic systems are not universally negative. Algorithms can facilitate access to supportive communities, mental health resources, and educational content, particularly for individuals who may lack offline support networks.[xxvii] For example, social media platforms have enabled LGBTQ+ young people to access identity-affirming communities and peer support, helping to reduce feelings of isolation and stigma.[xxviii] During the COVID-19 pandemic, algorithmic systems were also used to promote guidance from organisations such as the World Health Organization and the NHS, alongside wellbeing and public health campaigns.[xxix] Research by Li et al. (2021) found that public health content promoting hope and response efficacy on TikTok generated higher engagement and was subsequently amplified through recommendation systems.[xxx] This highlights the importance of how algorithms are designed and utilised, demonstrating that the same systems capable of amplifying harmful content can also be leveraged to promote credible health information, positive behaviours, and supportive messaging when aligned with public health objectives rather than solely engagement and profit.

AI and algorithmic systems are also increasingly used to support safeguarding and public health interventions online. AI moderation tools are used to detect and remove harmful content, identify cyberbullying, and flag potentially exploitative material.[xxxi] In some cases, machine learning systems have been developed to identify language patterns associated with suicidal ideation or emotional distress, allowing platforms to intervene with support prompts or wellbeing resources.[xxxii] These examples demonstrate that AI itself is not inherently harmful; rather, the risks arise from how algorithms are currently designed and optimised. While these technologies have the potential to prioritise safety, wellbeing, and credible information, existing platform models continue to largely prioritise engagement and advertising revenue, limiting the extent to which these protective capabilities are fully realised.

More Robust, Up to Date Regulation is needed

The regulatory environment has struggled to keep pace with the rapid evolution of AI technologies, leaving substantial gaps in oversight regarding the type of content that can be amplified and the speed at which it is disseminated. In the UK, legislative developments such as the Online Safety Act represent a step towards platform accountability, yet many provisions remain recent, reactive, or limited in scope. Notably, only in recent amendments has it been made explicitly illegal to create non-consensual sexually explicit images using AI, highlighting how regulation often follows, rather than anticipates, technological harm.[xxxiii]

The consequences of this regulatory lag are reflected in both individual case studies and broader research evidence. High-profile cases have demonstrated how young users can be rapidly exposed to harmful content through algorithmic recommendation systems, including material related to self-harm and suicide, contributing to significant mental health deterioration. High-profile cases have demonstrated how young users can be rapidly exposed to harmful content through algorithmic recommendation systems, including dangerous viral challenges, self-harm material, and extremist content. One notable example is the so-called “Blackout Challenge” circulated on TikTok, in which children were encouraged to intentionally restrict their breathing until losing consciousness.[xxxiv] The parents of four British teenagers are seeking to sue TikTok for the deaths of their children who are believed to have participated in this challenge. The US-based Social Media Victims Law Center claims that the deaths were “the foreseeable result of ByteDance’s engineered addiction-by-design and programming decisions”, which are “aimed at pushing children into maximising their engagement with TikTok by any means necessary.” [xxxv] Included in this legal challenge are the parents of 14-year-old Jools Sweeney, who died in April 2022 after reportedly participating in the “Blackout Challenge”, believed to have been encountered through social media. Following his death, his mother has campaigned for legislative reform after being unable to access his social media accounts to understand the circumstances surrounding his online activity. This resulted in a petition submitted to Parliament calling for bereaved parents to have greater access to their children’s social media data following their deaths.[xxxvi] Together, these cases illustrate the potentially severe consequences of limited platform transparency and insufficient regulation surrounding algorithmically amplified content.

Systematic reviews have similarly found consistent associations between social media use and increased risks of anxiety, depression, and poor body image among adolescents, with algorithmic amplification identified as a key intensifying factor. A systematic review across 13 studies on the impact of social media on young people by Keles et al. (2020) found that depression was the most commonly measured outcome.[xxxvii] Additionally, a systematic review examining social networking sites and body image identified that repeated exposure to idealised appearance-focused content contributed to increased body dissatisfaction, particularly among young women.[xxxviii] Broader population research has also shown that adolescents spending more than three hours per day on social media may face double the risk of poor mental health outcomes compared to peers with lower usage levels.[xxxix]

AI Chatbots

A key area where concerns surrounding AI and regulation are continuing to emerge is the increasing use of AI chatbots by children and young people, particularly within the context of growing demand for mental health support and limited access to timely services. NHS mental health waiting lists continue to increase. With approximately 1 in 5 young people in England aged between 8 to 25 having a “probable mental disorder”,[xl] many young people are turning towards AI-driven conversational tools for emotional support, advice, and companionship.[xli] While AI chatbots may provide immediate and accessible support for individuals experiencing loneliness, anxiety, or social isolation, concerns have emerged regarding the safety and appropriateness of these systems for vulnerable young users.

Alongside these concerns, AI chatbots have also demonstrated potential therapeutic benefits. AI-assisted interventions are increasingly being explored as a means of supplementing traditional mental health services, particularly where demand exceeds capacity. One emerging example is avatar therapy, a psychological intervention originally developed for individuals experiencing distressing auditory hallucinations, whereby patients interact with a computer-generated representation of the voice or source of distress under clinical supervision.[xlii] Recent developments have incorporated AI to create more responsive and realistic avatars, with early studies suggesting potential benefits for reducing distress and improving emotional wellbeing.[xliii] Similarly, AI companions are increasingly being used to provide emotional support, conversation, and social connection, particularly among individuals experiencing loneliness or isolation.[xliv] For some users, these systems may offer an accessible and non-judgemental space to discuss concerns, practise coping strategies, or access psychoeducational information. As NHS mental health services face increasing demand, such technologies may provide opportunities to complement existing support pathways and improve access to early intervention.

However, emerging research suggests that emotionally dependent relationships with AI companions may negatively affect wellbeing, particularly among individuals with limited offline social support. Zhang et al. (2025) found that users with fewer human relationships were more likely to engage heavily with AI companions, with frequent emotional disclosure associated with poorer wellbeing outcomes.[xlv] Similarly, Fang et al. (2025) identified that while voice-based chatbot interaction could modestly reduce loneliness, heavy daily use was associated with increased dependence and reduced real-world social interaction.[xlvi] Concerns have also been raised regarding the responses generated by AI systems in relation to serious mental health symptoms. Research has demonstrated that some chatbots may reinforce harmful behaviours or provide unsafe information when responding to disclosures of suicidal ideation or self-harm.[xlvii] These risks were reflected in the case of Adam Raine, a 16-year-old who died by suicide in April 2025 following months of conversations with ChatGPT. Court filings alleged that the chatbot failed to appropriately escalate disclosures of suicidal intent and, in some instances, provided harmful guidance relating to self-harm. His parents later initiated legal action against OpenAI, accusing the company of prioritising engagement over user safety.[xlviii] Together, these concerns demonstrate how AI systems designed to maximise engagement and emotional interaction may create new forms of risk for young people, particularly in the absence of effective regulation, safeguarding, and clinical oversight.

While much of the discussion surrounding AI and young people's mental health has focused on social media and digital interactions, concerns are also emerging regarding the impact of AI on employment opportunities and future career prospects. AI technologies are increasingly being used throughout recruitment processes to screen applications, assess candidates, and automate elements of hiring decisions.[xlix] Whilst this can speed up processes like resume screening and sourcing candidates, concerns have been raised that these systems may reproduce and amplify existing inequalities, as AI models are trained on historical data that can reflect longstanding social and structural biases.[l] A well-documented example is the COMPAS algorithm used in US courts, which was found to disproportionately misclassify Black defendants as higher risk compared to white defendants, illustrating how biased training data can translate into discriminatory outcomes within automated decision-making systems.[li] At the same time, advances in generative AI have raised concerns about the displacement of entry-level and graduate roles traditionally used by young people to gain experience and enter the workforce. Recent UK Government analysis has identified sharp declines in junior-level hiring across roles such as accounting, graphic design, software engineering, and data analysis, despite no comparable decline across senior roles.[lii] While the report emphasises that these trends cannot be directly attributed to AI, it notes that they are occurring in occupations where AI tools have rapidly advanced, raising early concerns about the potential impact of AI on entry-level graduate opportunities. For a generation already facing economic uncertainty, rising housing costs, and increased competition for employment, concerns regarding job security and future career opportunities may contribute to heightened levels of stress, anxiety, and reduced optimism about the future, with 59% of young people worried about the impact of AI on their future job security.[liii] While the long-term labour market impacts of AI remain uncertain, the perception that technology may limit opportunities for education-to-employment transitions represents another pathway through which AI may influence the mental wellbeing of young people.

Solutions

Efforts to mitigate the risks associated with social media have increasingly turned to restrictive approaches, including proposals to ban or significantly limit access for under-16s. However, emerging evidence suggests that outright bans are unlikely to be effective in isolation. Recent policy developments in Australia, for example, have highlighted the practical challenges of enforcement, including age verification, circumvention, and the displacement of young users to less regulated platforms, with more than 60% of Australian children still using social media despite the ban.[liv] More broadly, research indicates that prohibitive measures may fail to address the underlying drivers of harm, particularly where social media remains embedded in young people’s social and cultural lives.[lv][lvi]  As such, while age-based restrictions may form part of a wider strategy, they are insufficient as a standalone solution to the complex risks posed by AI-driven environments.

A more targeted and sustainable approach lies in holding social media companies accountable for the design and impact of their algorithmic systems. This includes greater regulatory pressure to ensure that platforms do not promote or profit from harmful content, particularly for younger users, and to require the adaptation of algorithms for under-16 accounts. One proposed alternative is to limit or remove algorithmic recommendation systems entirely for younger users, reverting to content based solely on accounts they actively follow. While this may reduce exposure to harmful amplification, it does not eliminate pre-existing risks associated with social media use, such as social comparison and harmful peer dynamics.

Alongside stronger regulation, there is also potential to redesign AI systems in ways that actively promote children and young people’s wellbeing. Rather than optimising solely for engagement, platforms could be required to adopt “wellbeing-by-design” approaches for under-18 accounts, where algorithms prioritise age-appropriate, educational, and protective content over emotionally provocative material. This could include reducing exposure to appearance-focused recommendations, limiting repetitive harmful content loops, and increasing algorithmic signposting towards mental health support services, crisis resources, and evidence-based public health information. AI systems could also be used proactively to identify patterns associated with distress, cyberbullying, self-harm, or exploitative behaviour, enabling earlier safeguarding interventions and support prompts.

There is also an opportunity for algorithms to support healthier digital behaviours among young people. Features such as automatic screen-time reminders, sleep-mode restrictions during night-time hours, and prompts encouraging breaks from prolonged scrolling could help reduce compulsive usage patterns linked to disrupted sleep and poorer wellbeing. Platforms could also utilise recommendation systems to promote physical activity, positive social interaction, and wellbeing campaigns, particularly during periods of isolation or increased mental health vulnerability. However, for these interventions to be effective, wellbeing and safety must become core objectives within platform design, rather than secondary considerations behind user engagement and advertising revenue.

Similar safeguards should also be applied to the growing use of AI chatbots for mental health support. Rather than positioning chatbots as substitutes for professional care, regulatory frameworks should require that they operate as complementary tools with clear limits on their role and capabilities. This could include mandatory escalation protocols for users disclosing self-harm, suicidal ideation, or abuse, alongside automatic signposting to crisis services and qualified mental health professionals. Greater transparency regarding how chatbot responses are generated, age-appropriate design standards, and independent safety testing for systems marketed towards young people may also help reduce potential harms. While AI chatbots have the potential to improve access to information, emotional support, and early intervention, these benefits are unlikely to be realised safely without robust oversight, clinical input, and ongoing evaluation of their impact on vulnerable users.

A robust approach to regulation and oversight needs to include sanctions if, for instance, a social media or AI company refuses to pay fines levied for breach of regulations, as well as ways to limit social media companies unnecessarily dragging out the process through appeals which have no legal merit. This might ultimately include denying access to the UK.

Alongside regulatory approaches for social media use, there is also a role for education systems in preparing young people for an AI-enabled labour market. Developing digital and AI literacy skills, including how to effectively use generative AI tools in professional contexts such as research, writing, data analysis, and idea generation, may enhance employability and reduce inequalities in access to emerging technologies. Embedding this within school curricula and early career training could help ensure that young people are not only protected from the risks of AI, but are also equipped to engage with it productively in the workplace.

Ultimately, the issue is not the existence of AI itself, but the objectives it is designed to serve. With appropriate safeguards, AI could be repurposed to identify and filter harmful content, promote credible information, and support user wellbeing. Achieving this, however, requires robust and enforceable regulation, building on frameworks such as the Online Safety Act, to ensure that technological innovation is aligned with the protection of children and young people, rather than driven solely by engagement and profit.

Conclusion

AI has the capacity to either amplify harm or promote wellbeing, depending on how it is designed and governed. The challenge is therefore not whether these technologies can be made safer, but whether regulators and technology companies are willing to prioritise wellbeing alongside innovation and profit. If social media is here to stay, then too must be the responsibility to make it safe.

Caitlin Gilbert    May 2026

 

[i] Mcfarlane, G. (2022). How Facebook, Twitter, Social Media Make Money From You. [online] Investopedia. Available at: https://www.investopedia.com/stock-analysis/032114/how-facebook-twitter-social-media-make-money-you-twtr-lnkd-fb-goog.aspx.

[ii] ‌Metzler, H. and Garcia, D. (2023). Social Drivers and Algorithmic Mechanisms on Digital Media. Perspectives on Psychological Science, [online] 19(5), pp.735–748.

[iii] Meta and Google found liable in social media addiction trial. (2026). BBC News. [online] 25 Mar. Available at: https://www.bbc.co.uk/news/articles/c747x7gz249o.

[iv] ‌Ofcom (2024). Children and Parents: Media Use and Attitudes Report. [online] Ofcom. Available at: https://www.ofcom.org.uk/siteassets/resources/documents/research-and-data/media-literacy-research/children/children-media-use-and-attitudes-2024/childrens-media-literacy-report-2024.pdf?v=368229.

[v] Martin, E.R. (2026). Social Media and Eating Disorders: A Growing Concern for Youth | ANAD - National Association of Anorexia Nervosa and Associated Disorders. [online] ANAD - National Association of Anorexia Nervosa and Associated Disorders. Available at: https://anad.org/social-media-and-eating-disorders-a-growing-concern-for-youth-2/.

[vi] De, D. El Jamal et al, (2025)

[vii] Legislation.Gov.UK (2023). Online Safety Act 2023. [online] Legislation.gov.uk. Available at: https://www.legislation.gov.uk/ukpga/2023/50.

[viii] Piatrov, I. (2024). Development of Adaptation of Machine Learning and Artificial Intelligence Into Facebook Algorithm. MARKETING IDENTITY, 562. Available at: https://www.researchgate.net/profile/Richard-Fedorko/publication/388115933_Exploring_the_Link_between_Positive_Online_Shopping_Experiences_and_Online_Shopping_Frequency_Insights_for_Mobile_Marketing_Strategies/links/67a5ff43645ef274a4742d38/Exploring-the-Link-between-Positive-Online-Shopping-Experiences-and-Online-Shopping-Frequency-Insights-for-Mobile-Marketing-Strategies.pdf#page=563

[ix] Psychology Today. (2026). Navigating the Impacts of AI-Driven Social Media Algorithms. [online] Available at: https://www.psychologytoday.com/gb/blog/the-modern-child/202601/navigating-the-impacts-of-ai-driven-social-media-algorithms.

[x] Dekker, C.A., Baumgartner, S.E. and Sumter, S.R. (2025). For you vs. for everyone: The effectiveness of algorithmic personalization in driving social media engagement. Telematics and Informatics, [online] 101, p.102300.

[xi] ‌Beyari, H. and Hashem, T. (2025). The Role of Artificial Intelligence in Personalizing Social Media Marketing Strategies for Enhanced Customer Experience. Behavioral Sciences, [online] 15(5), pp.700–700.

[xii] ‌eSafety Commissioner. (2025). An unfair fight – how algorithms are shaping our adolescents | eSafety Commissioner. [online] Available at: https://www.esafety.gov.au/newsroom/blogs/an-unfair-fight-how-algorithms-are-shaping-our-adolescents.

[xiii] brandography (2025). Social Media Algorithms and Mental Health | ACP. [online] Associated Clinic of Psychology. Available at: https://acp-mn.com/about-acp/blog/social-media-algorithms-and-mental-health/.

[xiv] ‌Collins, D. (2022). Clerk to the Senior Coroner REGULATION 28 REPORT TO PREVENT FUTURE DEATHS THIS REPORT IS BEING SENT TO: Secretary of State for Digital, Culture, Media and Sport 4th Floor 100 Parliament Street London SW1A 2BQ. [online] Available at: https://www.judiciary.uk/wp-content/uploads/2022/10/Molly-Russell-Prevention-of-future-deaths-report-2022-0315_Published.pdf.

[xv] ‌Naughton, J. (2022). Molly Russell was trapped by the cruel algorithms of Pinterest and Instagram. The Observer. [online] 1 Oct. Available at: https://www.theguardian.com/commentisfree/2022/oct/01/molly-russell-was-trapped-by-the-cruel-algorithms-of-pinterest-and-instagram.

[xvi] Ahmmad, M., Shahzad, K., Iqbal, A. and Latif, M. (2025). Trap of Social Media Algorithms: A Systematic Review of Research on Filter Bubbles, Echo Chambers, and Their Impact on Youth. Societies, [online] 15(11), p.301. doi:https://doi.org/10.3390/soc15110301.

[xvii] ‌Christakis, D.A. and Hale, L. (2025). Toward Defining Problematic Media Usage Patterns in Adolescents. JAMA. [online] doi:https://doi.org/10.1001/jama.2025.6113.

[xviii] Behera, N., Khuntia, S., Pandey, K. and Shankar, S. (2025). Impact of Social Media Use on Physical, Mental, Social, and Emotional Health, Sleep Quality, Body Image, and Mood: Evidence from 21 Countries—A Systematic Literature Review with Narrative Synthesis. International Journal of Behavioral Medicine. doi:https://doi.org/10.1007/s12529-025-10411-9.

[xix] Ofcom (2023). Children’s Media Lives. [online] Available at: https://www.ofcom.org.uk/media-use-and-attitudes/media-habits-children/childrens-media-lives.

[xx] ‌Gomes, K. and Goldman, R.D. (2024). Screen time and sleep in children. Canadian family physician/Le médecin de famille canadien, 70(6), pp.388–390. doi:https://doi.org/10.46747/cfp.7006388.

[xxi] King’s College London (2016). Bedtime use of media devices doubles risk of poor sleep in children | Website archive | King’s College London. [online] Available at: https://www.kcl.ac.uk/archive/news/ioppn/records/2016/october/bedtime-use-of-media-devices-doubles-risk-of-poor-sleep-in-children.

[xxii] Liu, J., Ji, X., Pitt, S., Wang, G., Rovit, E., Lipman, T. and Jiang, F. (2022). Childhood sleep: physical, cognitive, and behavioral consequences and implications. World Journal of Pediatrics, [online] 20(2). doi:https://doi.org/10.1007/s12519-022-00647-w.

[xxiii] Sanzari, C.M., Gorrell, S., Anderson, L.M., Reilly, E.E., Niemiec, M.A., Orloff, N.C., Anderson, D.A. and Hormes, J.M. (2023). The impact of social media use on body image and disordered eating behaviors: Content matters more than duration of exposure. Eating Behaviors, [online] 49(101722), p.101722. doi:https://doi.org/10.1016/j.eatbeh.2023.101722.

[xxiv] ‌ProQuest (2025). Identity, the Manosphere, and Misogyny: Unpacking the Digital Pipeline - ProQuest. [online] Available at: https://search.proquest.com/openview/b61c8c31cce1cd9852d43ccc8b2a24b1/1?pq-origsite=gscholar&cbl=18750&diss=y [Accessed 12 Jun. 2026].

[xxv] APA (2026). APA PsycNet. [online] Available at: https://psycnet.apa.org/record/2027-06772-041.

[xxvi] ‌Zhang, Y., Zhao, D., Hancock, J.T., Kraut, R. and Yang, D. (2025). The Rise of AI Companions: How Human-Chatbot Relationships Influence Well-Being. [online] arXiv.org. Available at: https://arxiv.org/abs/2506.12605.

[xxvii] URMC Newsroom. (2025). The Positives of Social Media for Teens and How Parents Can Guide Safe Use. [online] Available at: https://www.urmc.rochester.edu/news/publications/health-matters/the-positives-of-social-media-for-teens-and-how-parents-can-guide-safe-use.

[xxviii] Berger, M.N., Taba, M., Marino, J.L., Lim, M.S.C. and Skinner, S.R. (2022). Social Media Use and Health and Well-being of Lesbian, Gay, Bisexual, Transgender, and Queer Youth: Systematic Review. Journal of Medical Internet Research, [online] 24(9). doi:https://doi.org/10.2196/38449.

[xxix] Ramsey, N., Obeidallah, M. and Abraham, A. (2023). Impact of COVID-19 on adolescent health and use of social media. Current Opinion in Pediatrics, [online] 35(3). doi:https://doi.org/10.1097/mop.0000000000001248.

[xxx] Li, Y., Guan, M., Hammond, P. and Berrey, L.E. (2021). Communicating COVID-19 information on TikTok: a content analysis of TikTok videos from official accounts featured in the COVID-19 information hub. Health Education Research, 36(3), pp.261–271. doi:https://doi.org/10.1093/her/cyab010.

[xxxi] Scott, C., Marcu, G., Anderson, R., Newman, M. and Schoenebeck, S. (2023). Trauma-Informed Social Media: Towards Solutions for Reducing and Healing Online Harm. Trauma-Informed Social Media: Towards Solutions for Reducing and Healing Online Harm, [online] (23). doi:https://doi.org/10.1145/3544548.3581512.

[xxxii] ‌Thakkar, A., Gupta, A. and De Sousa, A. (2024). Artificial intelligence in positive mental health: a narrative review. Frontiers in digital health, [online] 6(1280235). doi:https://doi.org/10.3389/fdgth.2024.1280235.

[xxxiii] Home Office (2026). Government leads global fight against deepfake threats. [online] GOV.UK. Available at: https://www.gov.uk/government/news/government-leads-global-fight-against-deepfake-threats.

[xxxiv] Felbin, S. and Talbert, S. (2023). Another Death From The ‘Blackout Challenge’ Has Led To A TikTok Lawsuit. [online] Women’s Health. Available at: https://www.womenshealthmag.com/health/a38603617/blackout-challenge-tiktok-2021/.

[xxxv] ‌Fraser, G. (2025). TikTok sued by parents of UK teens after alleged challenge deaths. BBC News. [online] 7 Feb. Available at: https://www.bbc.co.uk/news/articles/c0lz2x60w46o.

[xxxvi] Joolslaw. (2024). Joolslaw. [online] Available at: https://www.joolslaw.co.uk/.

[xxxvii] Keles, B., McCrae, N. and Grealish, A. (2020). A Systematic review: the Influence of Social Media on depression, Anxiety and Psychological Distress in Adolescents. International Journal of Adolescence and Youth, 25(1), pp.79–93. doi:https://doi.org/10.1080/02673843.2019.1590851.

[xxxviii] Fioravanti, G., Bocci Benucci, S., Ceragioli, G. and Casale, S. (2022). How the Exposure to Beauty Ideals on Social Networking Sites Influences Body Image: A Systematic Review of Experimental Studies. Adolescent Research Review, [online] 7(7), pp.419–458. doi:https://doi.org/10.1007/s40894-022-00179-4.

[xxxix] Riehm, K.E., Feder, K.A. and Tormohlen, K.N. (2019). Associations between Time Spent Using Social Media and Internalizing and Externalizing Problems among US Youth. JAMA Psychiatry, [online] 76(12), pp.1266–1273. doi:https://doi.org/10.1001/jamapsychiatry.2019.2325.

[xl] NHS Digital (2023). Mental health of children and young people in england, 2023 - wave 4 follow up to the 2017 survey. [online] NHS Digital. Available at: https://digital.nhs.uk/data-and-information/publications/statistical/mental-health-of-children-and-young-people-in-england/2023-wave-4-follow-up.

[xli] ‌Atske, S. (2025). Teens, Social Media and AI Chatbots 2025. [online] Pew Research Center. Available at: https://www.pewresearch.org/internet/2025/12/09/teens-social-media-and-ai-chatbots-2025/.

[xlii] Garety, P.A., et al., (2024). Digital AVATAR therapy for distressing voices in psychosis: the phase 2/3 AVATAR2 trial. PubMed. doi:https://doi.org/10.1038/s41591-024-03252-8.

[xliii] Medica-tradefair.com. (2026). MEDICA. [online] Available at: https://www.medica-tradefair.com/en/media-news/spheres-of-medica-magazine/digital-health/virtual-avatars-anxiety-vr-therapy [Accessed 12 Jun. 2026].

[xliv] Merrill, K., Mikkilineni, S.D. and Dehnert, M. (2025). Artificial intelligence chatbots as a source of virtual social support: Implications for loneliness and anxiety management. Annals of the New York Academy of Sciences, 1549(1), pp.148–159. doi:https://doi.org/10.1111/nyas.15400

[xlv] ‌Liu, J., Wang, X. and Zhang, J. (2025). Investigating Elderly Individuals’ Acceptance of Artificial Intelligence (AI)-Powered Companion Robots: The Influence of Individual Characteristics. Behavioral Sciences, [online] 15(5), pp.697–697. doi:https://doi.org/10.3390/bs15050697.

[xlvi] ‌Fang, C., Liu, A., Danry, V., Lee, E., Chan, S., Pataranutaporn, P., Maes, P., Phang, J., Lampe, M., Ahmad, L. and Agarwal, S. (n.d.). How AI and Human Behaviors Shape Psychosocial Effects of Extended Chatbot Use: A Longitudinal Randomized Controlled Study. [online] Available at: https://arxiv.org/pdf/2503.17473.

[xlvii] ‌Wells, S. (2025). Exploring the Dangers of AI in Mental Health Care | Stanford HAI. [online] Stanford.edu. Available at: https://hai.stanford.edu/news/exploring-the-dangers-of-ai-in-mental-health-care.

[xlviii] ‌Yousif, N. (2025). Parents of teenager who took his own life sue OpenAI. BBC News. [online] 27 Aug. Available at: https://www.bbc.co.uk/news/articles/cgerwp7rdlvo.

[xlix] ‌Hosain, M.S., Amin, M.B., Debnath, G.C. and Rahaman, M.A. (2025). The use of Artificial Intelligence (AI) in the hiring process: Job applicants’ perceptions of procedural justice. Computers in Human Behavior Reports, 19, p.100713. doi:https://doi.org/10.1016/j.chbr.2025.100713.

[l] Yohannis, A., and Kolovos, D.S. (2022). Towards model-based bias mitigation in machine learning. doi:https://doi.org/10.1145/3550355.3552401.

[li] Larson, J., Mattu, S., Kirchner, L. and Angwin, J. (2016). How We Analyzed the COMPAS Recidivism Algorithm. [online] ProPublica. Available at: https://www.propublica.org/article/how-we-analyzed-the-compas-recidivism-algorithm.

[lii] ‌Gov.UK (2026). A snapshot of entry-level hiring in the UK. [online] GOV.UK. Available at: https://www.gov.uk/government/publications/entry-level-hiring-in-the-uk-a-snapshot/a-snapshot-of-entry-level-hiring-in-the-uk.

[liii] Hill, A. (2026). Most young adults in UK are anxious about jobs and the economy, research suggests. [online] the Guardian. Available at: https://www.theguardian.com/society/2026/jan/27/uk-young-people-anxious-jobs-economy.

[liv] Hill, G. (2026). More than 60% of Australian children still using social media despite ban for under-16s, research shows | Molly Rose Foundation. [online] Molly Rose Foundation. Available at: https://mollyrosefoundation.org/more-than-60-of-australian-children-still-using-social-media-despite-ban-for-under-16s-research-shows/.

[lv] J. Fatt, S. and Fardouly, J. (2025). Debate: Social media in children and young people – time for a ban? Weighing up the implications and limitations of age‐based social media restrictions. Child and Adolescent Mental Health. doi:https://doi.org/10.1111/camh.70034.

[lvi] McAlister, K., Beatty, C., Smith-Caswell, J., Yourell, J. and Huberty, J. (2024). Social Media Use in Adolescents: Bans, Benefits, and Emotion Regulation Behaviors. JMIR Mental Health, [online] 11(11). doi:https://doi.org/10.2196/64626.