Written by Julien Ricciarelli-Bonnal

7 August 2026

The Essentials

A New Mexico court has ordered Meta to pay $567 million into a fund dedicated to adolescent mental health and to introduce changes to Facebook and Instagram over the next five years. The ruling matters because it does not focus only on content posted by users, but also on how the platforms themselves are designed: notifications, screen time, addictive mechanics, protections for minors and certain AI tools. Meta has said it will appeal. Behind the spectacular figure lies a much broader issue: can social networks continue to present harm to young users as a consequence of individual behaviour when courts are beginning to examine the design choices that shape that behaviour? 

For years, major platforms have relied on a relatively comfortable distinction: they host content, but they do not create all of it, and therefore cannot be held responsible for every harmful post or every individual behaviour that occurs through their services. In the United States, that distinction has been reinforced by Section 230 of the Communications Decency Act, which provides online platforms with significant protection from certain liabilities linked to third-party content.

The ruling issued in New Mexico on 6 August approaches the problem differently. Judge Bryan Biedscheid concluded that the case was not solely about what users publish on Facebook or Instagram, but also about the way Meta designed its products, particularly the mechanisms used to retain attention, manage notifications, protect minors and operate certain artificial-intelligence features. The court therefore rejected Meta’s argument that Section 230 was sufficient to shield the company in this case. 

Meta must now pay $567 million into a fund intended to support adolescent mental-health treatment, prevention and screening, while a five-year decree requires the company to introduce or strengthen a range of youth-safety measures. The group disputes the ruling and has announced its intention to appeal. 

The amount is striking. But for marketing, digital products and the wider platform economy, the potential precedent may matter even more than the money.

The Case Is No Longer Only About What Users Post

Social networks have always had an interest in separating their own design choices from the behaviour of the people who use them. A violent video, a fraudulent message or an abusive post can be attributed to the person who published it, leaving the platform in the role of intermediary, with moderation duties that vary according to jurisdiction and circumstance.

The New Mexico case shifts the centre of gravity. The state argued that certain characteristics of the products themselves could contribute to harm among young users, particularly when systems are designed to increase time spent, encourage repeated returns or keep people moving through continuous loops of content.

That distinction is fundamental. When a court examines a piece of user-generated content, the platform can argue that it did not create it; when it examines a notification system, recommendation mechanism, screen-time control or protection aimed at minors, the discussion moves much closer to decisions made directly by the company itself.

The judge followed that reasoning in significant parts of the ruling, treating several claims as questions of product design rather than traditional editorial liability for third-party speech. At the same time, some remedies requested by the state were rejected on First Amendment or technological-feasibility grounds, which means the decision did not simply accept every allegation made against Meta. 

A Notification Is Never Entirely Neutral

A notification looks trivial. It tells a user that someone reacted to a post, that new content is available or that something on the platform deserves attention, yet multiplied across millions of users and optimised to increase return visits, these small prompts become part of the core machinery of the attention economy.

Platforms have spent years refining such mechanisms, testing timing, wording, frequency and behavioural impact in much the same way an ecommerce business optimises a checkout funnel. There is nothing inherently suspicious about improving a digital product, but the problem becomes more difficult when the user is a teenager and the objective of increasing engagement begins to conflict with the need to limit excessive use.

A feature can be excellent at increasing time spent while being far less desirable from a wellbeing perspective. The tension therefore moves away from abstract moral arguments and into the metrics themselves: what exactly is the company trying to optimise, and at what point does a successful engagement mechanism become a liability?

The issue is no longer simply whether teenagers spend too much time on Instagram. It is also how much effort Instagram devotes to bringing them back, and whether those efforts should be judged differently when the audience is made up of minors.

That is what makes the problem considerably harder for Meta and for the sector more broadly. A platform can introduce parental controls and wellbeing tools while still operating a product whose commercial success is tied to repeated use, meaning youth protection cannot remain a decorative layer added on top of systems optimised in the opposite direction.

Mental Health Is Becoming a Product-Design Issue

Debates about social media and young people have long focused on content: social comparison, bullying, sexualisation, violence, misinformation or exposure to harmful communities. Those questions remain important, but the New Mexico ruling expands the discussion towards the architecture of the product itself.

The court ordered Meta to introduce measures relating to screen time, notifications, user information and safeguards around certain AI chatbot functions. The state had also sought wider remedies involving age verification, default protections and measures against the sexual exploitation of minors. 

This approach brings platforms closer to other industries in which product design can be examined when it increases risk. A company cannot always defend itself simply by arguing that consumers could have behaved differently if some of the product’s own characteristics were deliberately designed to shape that behaviour.

The comparison must nevertheless remain cautious. The relationship between social-media use and mental health is complex, differs significantly between individuals and cannot be reduced to a single causal mechanism. The New Mexico judgment does not establish that every hour spent on Instagram causes psychological harm; it establishes that, in this case, a judge considered certain design decisions by Meta legally relevant enough to justify substantial remedies.

That distinction will probably sit at the heart of the appeal. Meta can challenge the legal basis of the decision, the scope of the remedies and the way the court connects platform design with the harms alleged by the state.

Meta Now Has to Defend What It Once Treated as Product Choices

The difficulty for Meta is that the features under scrutiny are not peripheral. Recommendations, notifications, personalisation and engagement mechanisms are precisely what have made modern social platforms commercially powerful, because a free service generates more advertising opportunities when users return frequently and consume more content.

Limiting screen time or reducing certain prompts is therefore not merely a safety decision. It can come into direct tension with the economic logic of a product whose value is partly built on attention, which makes the discussion considerably more uncomfortable than a conventional argument about removing harmful posts.

Meta has already introduced numerous protections for teenagers, from account restrictions to parental controls and changes in the way certain content is recommended, and the company regularly highlights the resources it devotes to safety. The New Mexico ruling does not say those efforts do not exist; it concludes that they were insufficient to address the problems established in this particular case. 

To design a digital experience without allowing engagement to become the only measure of success⁠, companies increasingly need to consider the reputational, regulatory and social consequences of product decisions much earlier. A mechanism that improves a metric today can become the central exhibit in a lawsuit several years later.

A Ruling That Could Reach Far Beyond New Mexico

The judgment directly concerns New Mexico, and Meta’s appeal means its final legal significance remains uncertain. It nevertheless arrives at a moment when major social platforms are facing a growing number of US lawsuits concerning young users, mental health and the design of their products, which makes the reasoning behind this case particularly important.

If this approach survives appeal, social-media companies may have to think very differently about features aimed at younger audiences. It would no longer be enough to demonstrate that certain harmful posts are removed; companies could increasingly be expected to show that interfaces, notification systems, recommendation engines and engagement mechanics were themselves designed with meaningful safeguards.

The ruling may also change the language used around engagement. For years, unusually high engagement was presented almost entirely as evidence that a product was relevant and satisfying to users. When the user is a minor, exceptionally strong engagement can now be interpreted differently, not only as evidence of interest but potentially as evidence that the product has become too effective at holding attention.

The implications therefore extend well beyond Meta. TikTok, YouTube, Snap and every platform built around similar engagement mechanics have reason to watch what happens next, because the underlying question is much broader than Instagram: how far can a digital product be optimised to retain attention before that optimisation itself becomes a legal risk?

Meta will try to overturn or narrow the decision, and the case is far from over. But New Mexico has already shifted an important boundary: platforms may increasingly be judged not only on what they allow users to publish, but also on what they deliberately build to keep them from leaving.

We support companies that want to design marketing and digital experiences without sacrificing their reputation to engagement alone⁠.