Meta’s $18 Billion Settlement Is a Warning. Young People Still Need to Know How the Machine Works

When Meta agreed this week to pay up to US$18 billion to settle US state claims concerning harm to children, much of the reporting quite reasonably concentrated on the new protections.

Teen accounts on Facebook and Instagram are to face a default two-hour daily limit. Overnight access will be blocked between midnight and 6 am. Push notifications will be disabled during school hours. Likes and reactions will be hidden by default. Age checks will become tougher. A non-personalised feed is also due to be introduced as part of the settlement.

Meta denies wrongdoing (Reuters, 26 August 2026).

These are substantial changes. They also tell us something uncomfortable.

For years, families were often given the impression that managing social media was mainly a matter of teaching children self-control. Put the phone away. Turn off notifications. Stop scrolling. Go outside.

Now some of those controls are being built into the platforms themselves.

That change matters because it moves at least part of the discussion away from the child and towards the design of the technology.

The harder question, however, survives the settlement.

Why did these systems become so difficult to leave in the first place?

That is the question I wanted young people to wrestle with when developing the Brain Rot! Learning System.

The phone is only the visible part. A recent commentary on the Meta case made the argument particularly sharply. The concern, it suggested, goes further than any single platform feature. Social media operates within a commercial model built around observing behaviour, learning from it and using those predictions to keep people engaged and make advertising more effective. The shorthand phrase “you are the product” is useful, but children deserve a better explanation than that.

Meta does not normally package up a teenager’s personal dossier and sell it across a counter to an advertiser. The system is more sophisticated. Platforms can collect signals from behaviour such as clicks, viewing time, searches, follows and other interactions. Those signals help build predictions about interests and behaviour. Advertisers then pay for the ability to reach particular kinds of audiences.

Attention has commercial value because attention creates opportunities to show advertising, gather further behavioural information and improve future predictions. Once a student understands that loop, features such as infinite scroll, recommendations, autoplay, streaks and notifications stop looking like random pieces of an app. They can be examined as design choices.

This is where the conversation with young teens becomes much more interesting.

Instead of asking only, “How many hours were you online?”, we can ask, “What was the system learning while you were there?” Lizzy Winstone and colleagues explored this territory in a 2026 BMC Public Health study of young people aged 14 to 19. Their interviews examined how adolescents experienced algorithmically curated Instagram and TikTok feeds. Participants described feeds that changed in response to their engagement and, in some cases, repeatedly served material connected with difficult emotions or interests. The authors use the term algorithmically structured exposure to describe part of this relationship between user behaviour, recommendations and commercial platform design. This was a qualitative study, so it does not establish that recommendation algorithms cause poor mental health. It does show why studying “screen time” alone misses a great deal of what young people actually encounter online. (Winstone, L., Parsonage, J., Cross, L., Muir, C., Randhawa, A., MacArthur, G., Garavini, E., & Asiedu-Addo, E., 2026).

That distinction is important.

A 2024 umbrella review by Andrea Sala, Lorenzo Porcaro and Emilia Gómez reached a similarly complicated conclusion after examining existing reviews of adolescent social-media use and wellbeing. Social media can offer benefits as well as risks, and outcomes vary according to the young person, the kind of use and the design of the platform. The authors specifically identify platform design and limited access to platform data as important parts of the research problem. (Sala, A., Porcaro, L., & Gómez, E., 2024).

Another large 2024 systematic review and meta-analysis, led by Oli Ahmed at the Australian National University, distinguished general social-media use from problematic social-media use. Across the studies they examined, problematic use had stronger associations with depression, anxiety, sleep problems and poorer wellbeing than ordinary measures of time spent using social media. Association is not causation, but the finding again warns against reducing the discussion to a simple count of hours. (Ahmed, O., Walsh, E. I., Dawel, A., Alateeq, K., Espinoza Oyarce, D. A., & Cherbuin, N., 2024).

Young people need enough knowledge to see this system for themselves.

That is the educational gap Brain Rot! is designed to occupy.

Story gives us somewhere safer to start

Adults can make discussions about persuasive technology surprisingly personal, very quickly.

• “How much time do you spend on TikTok?”

• “Why were you still online at midnight?”

• “What were you watching?”

For a thirteen-year-old, the shutters can come down before the useful conversation has begun. Story gives us another route. In Brain Rot!, students can look at Beam, Bindi, Chi, Mia and the fictional Zipp app first. They can watch what happens when a system begins learning what holds a character’s attention. They can look for evidence. They can argue about who has control. They can decide whether a particular feature is helpful, persuasive or manipulative.

Nobody has to begin by admitting what happened on their own phone last night.

The distance created by fiction is deliberate. Once students have pulled apart the fictional system, they have language for examining real ones.

The Brain Rot! resources offer several ways of doing that because not every child learns well from the same format.

• The multisensory Read-Aloud Kit uses selected passages, listening, discussion and visual prompts. A parent, librarian or teacher can stop at the moment a character makes a decision and investigate what influenced it.

• The 5-Day Mini-Unit Pack uses a curated sequence of chapters and activities. It is designed for families and educators who want something deeper than a single discussion but do not need the complete program.

• The 8-minute comic-book-style animation episode gives visual learners another entry point. The accompanying activities turn the episode into evidence: students identify persuasive features, clues that the system is learning from its users, warning signs and possible design alternatives.

• Then there is the illustrated Brain Rot! novella and Activity Guide, with more than 40 zero-tech and technology-integrated activities.

Some of my favourite activities deliberately remove the screen. Students can physically build a feedback loop. They can follow the movement of attention, behavioural data and advertising money. They can sort examples of signal and noise. They can examine what an app might infer from repeated behaviour. They can redesign a fictional platform so that its success depends less on keeping someone online for as long as possible. The absence of a device is useful. A recommendation system becomes much easier to question when its logic is sitting on the table in front of you as cards, arrows, tokens and choices.

Technology-supported activities can then take the next step: analysing, modelling, researching, designing and testing.

This is also why Brain Rot! sits comfortably within STEM. A recommendation system can be treated as a system with inputs, processing, outputs and feedback. Students can ask what data are collected, what pattern might be inferred, what prediction follows and what happens when the output changes the user’s next action. They are doing computational thinking while investigating something already present in their daily lives.

The current ISTE Standards make the same connection. The Digital Citizen standard asks students to safeguard their wellbeing and make informed choices online. The Computational Thinker standard includes data analysis, decomposition, models and algorithmic thinking. The Innovative Designer standard asks learners to develop and test solutions rather than accepting existing designs as inevitable. (ISTE)

For me, that last part is especially important. Young people should be able to criticise an app. They should also be able to imagine a better one.

Do teenagers actually understand profiling?

Not always.

Sonali Srivastava, Terhi-Anna Wilska and Jussi Nyrhinen held eight focus-group discussions with 38 Finnish adolescents aged 13 to 16. The researchers found considerable variation in what participants understood about online profiling and targeted advertising.

Some accepted tracking of their online behaviour as fairly normal. Others described profiling as invasive or “creepy”. Several had gaps in their understanding of how commercial data gathering worked. Some participants even said schools should teach more about how online data are commercially repurposed.

It was a small qualitative study in Finland, so its findings should not be treated as a universal portrait of teenagers. Its educational implication is still hard to ignore: knowing that advertising exists is different from understanding how behavioural profiling supports it. (Srivastava, S., Wilska, T.-A., & Nyrhinen, J., 2023).

That is the difference between teaching children to spot an advertisement and teaching them to ask why this advertisement, this video or this recommendation appeared in their feed at this moment.

Those questions belong in contemporary media literacy.

They also belong in conversations about wellbeing.

The strongest current evidence does not support a crude claim that social media damages every teenager. Young people use these platforms for friendship, entertainment, creativity, information and belonging. Effects vary greatly. Nor should education become a convenient way for technology companies to transfer responsibility back to children. A thirteen-year-old with a media-literacy lesson behind them is still a thirteen-year-old facing systems built by extremely well-resourced companies.

A 2025 scoping review from researchers at Orygen and colleagues examined 70 policy and platform recommendations concerning social media and youth mental health. The recurring themes were accountability, transparency, collaboration, safety by design and restrictions on young people's access. Significantly, the review concentrated on what governments and technology companies should do rather than telling children to manage the problem themselves. (Chhabra, J., Pilkington, V., Benakovic, R., Wilson, M. J., La Sala, L., & Seidler, Z., 2025).

Education therefore belongs beside safer design and regulation, not in place of them.

The Meta settlement is a good example. Legal pressure can change defaults, limits and platform architecture. Parents can set boundaries. Schools can establish sensible device policies. But a young person will eventually meet another platform, another recommendation system, another game, another AI companion or another persuasive interface.

A rule cannot travel everywhere with them.

Understanding can.

That is what I want students to take from Brain Rot!.

Not fear of technology. Not a lecture about phones. Not the vague instruction to “use screens responsibly”. I want them to be able to look beneath an interface and work out what is happening.

• What information am I giving this system?

• What can it infer from that information?

• Why does it want me to stay?

• Who earns money if I keep engaging?

• What is happening to my mood, attention, sleep or friendships while I do?

And perhaps the most useful question of all:

Does this technology work for me, or am I beginning to work for it?

References and further reading

Ahmed, O., Walsh, E. I., Dawel, A., Alateeq, K., Espinoza Oyarce, D. A., & Cherbuin, N. (2024). Social media use, mental health and sleep: A systematic review with meta-analyses. Journal of Affective Disorders, 367, 701–712. https://doi.org/10.1016/j.jad.2024.08.193 (PubMed)

Chhabra, J., Pilkington, V., Benakovic, R., Wilson, M. J., La Sala, L., & Seidler, Z. (2025). Social Media and Youth Mental Health: Scoping Review of Platform and Policy Recommendations. Journal of Medical Internet Research, 27, e72061. https://doi.org/10.2196/72061 (PubMed)

Sala, A., Porcaro, L., & Gómez, E. (2024). Social Media Use and adolescents' mental health and well-being: An umbrella review. Computers in Human Behavior Reports, 14, 100404. https://doi.org/10.1016/j.chbr.2024.100404 (ScienceDirect)

Schreurs, L., & Vandenbosch, L. (2026). Enhancing adolescents’ emotion regulation in the social media context: a cluster randomized controlled trial of the Vibe Check social media literacy intervention. Educational Psychology, 46(5), 726–744. https://doi.org/10.1080/01443410.2025.2593545 (DOI)

Srivastava, S., Wilska, T.-A., & Nyrhinen, J. (2023). Awareness of digital commercial profiling among adolescents in Finland and their perspectives on online targeted advertisements. Journal of Children and Media, 17(4), 559–578. https://doi.org/10.1080/17482798.2023.2257813 (Tandfonline)

Winstone, L., Parsonage, J., Cross, L., Muir, C., Randhawa, A., MacArthur, G., Garavini, E., & Asiedu-Addo, E. (2026). Algorithmic recommendation and adolescent mental health: a grounded theory study of social and commercial determinants. BMC Public Health. https://doi.org/10.1186/s12889-026-28300-5 (PubMed)

Reuters. (26 August 2026). What Meta agreed to in US teen safety settlement. This is a news source rather than peer-reviewed research and is used here for the terms of the current settlement. (Reuters)

International Society for Technology in Education. ISTE Standards for Students and Computational Thinking Competencies. These are educational standards documents rather than research studies. (ISTE)

Suggested hashtags: #MetaSettlement #Meta #InstagramSafety #SocialMediaSafety #OnlineSafety #YouthMentalHealth #DigitalWellbeing #DigitalCitizenship #MediaLiteracy #AlgorithmicLiteracy #PersuasiveTechnology #ParentingTeens #TeachersOfTeens #STEMEducation #EdTech

This version deliberately leaves a little more texture in the writing rather than making every paragraph and section match. I have also avoided treating the references as interchangeable evidence: the Meta facts come from current legal reporting, the wellbeing claims from reviews and empirical studies, and the educational alignment from ISTE itself.