A few days ago, the Los Angeles Superior Court issued a historic ruling: it ordered Meta and Alphabet to pay three million dollars in damages to a twenty-year-old girl for the psychological harm caused by Facebook and YouTube: the girl, who had been using social media since she was six, had begun experiencing suicidal thoughts already during childhood.
The California jury found that the platforms deliberately designed algorithms that create addiction – infinite scrolling, notifications, personalized recommendations aimed at prolonging the young woman’s time on the platform.
This ruling could have a huge impact on the laws regulating the activity of digital platforms, if parliaments manage to emerge from the stranglehold imposed by the lobbies defending Big Tech’s interests.
Observing the evolution of platforms over the last twenty years, the anomalous nature of the digital industry emerges.
In other sectors, companies are required to follow strict rules before marketing a product to prevent harm to citizens and the environment: this happens when a drug, a new car, a food product, or any technology that goes into consumers’ hands is put on the market.
In the case of the algorithms that regulate the functioning of social networks, which accompany the lives of billions of people, we are the guinea pigs: there are no rules to protect us from the potential harm caused by these technologies.
Yet there are hundreds of studies that for over a decade have documented the harm caused by social networks, and not only to the youngest.
WHAT WE KNOW
The California court did not discover anything new. I myself published a book in 2024 (The Human Farm, Treccani books) where I describe the social and political problems created by the digital industry. Let me cite the most significant ones before delving into the resistance of the platforms and possible solutions.
Jean Twenge, a psychologist at San Diego State University who has studied the issue for many years, claims to have noticed a sharp change in the emotional state of adolescents since 2011, when the percentage of depression and suicides among young people began to soar.
Twenge reveals that, in a document leaked from inside Facebook, the company highlights to advertisers its ability to determine the emotional state of very young users based on their behavior on the site, even identifying “moments when young people need a confidence boost.”
Psychologist Jonathan Haidt, in the bestseller The Anxious Generation (Rizzoli, 2024), argues that the surge in depression and self-harm among adolescents has been observed since 2012, the year when smartphone and social media use spread among young people, and that there is a strong correlation between the two. Andreas Schleicher, the inventor of the OECD’s PISA program, has long reported a decline in the quality of education and skills of adolescents starting precisely in 2012.
The Italian Society of Pediatrics, analyzing 68 studies conducted from 2004 to 2022, concludes that the more time young people spend on social networks, the higher the risk they develop depressive symptoms; in particular, a “significant association” is found between depression in young people under 18 and the use of platforms like Facebook, Instagram, and TikTok.
One of the researchers, Rino Agostiniani, says: “What emerges unequivocally is that the more time children and adolescents spend on digital devices, the higher levels of depression are reported. And this happens without major geographical distinctions: from Sweden to Egypt.”
The mechanism is now well known. The constant collection of personal data allows platforms to identify each person’s tastes and vulnerabilities.
The algorithms, unaware of the damage they may cause, deliver personalized content to each user in order to keep them glued to the screen to monetize every second of their attention.
It was Shoshana Zuboff (The Age of Surveillance Capitalism, Luiss University Press, 2019) who most effectively described the strategy of platforms that appropriate human experience and turn it into the (free) raw material that allows them to predict and modify user behavior, a valuable commodity for advertisers.
Zuboff calls this ability the “global architecture of behavior modification.”
THE HIDDEN GOAL
Now the California ruling makes this contradiction explicit, which dozens of legal scholars have been discussing for years.
The heart of the problem is simple: a drug, before being marketed, must undergo years of clinical trials, obtain approval from the EMA in Europe or the FDA in the United States, list side effects on packaging, and be monitored after market release.
A car must meet certified safety standards, pass crash tests, have approved seat belts, and so on. Every food product must comply with hygiene standards and declare ingredients on the label.
Conversely, an algorithm designed to maximize the psychological addiction of millions of children requires none of this. It can be freely launched on the market, enter the phones of billions of users, and will be judged – if ever it is – only retrospectively, in a courtroom, when it can be proven what damage has been caused.
Legal scholar Frank Pasquale already noted in 2015 (The Black Box Society) how large platforms, which wield enormous power over the lives of billions of people, operate through opaque systems that escape any external control. According to Pasquale, this opacity is a deliberate choice: the proprietary nature of algorithms limits understanding of their functioning, and the speed with which they evolve allows them to evade any legislative effort.
Those opposing excessive intrusion by public bodies point out that this is a slippery argument.
Algorithms can be opaque even to their own designers. Often machine learning systems are unable to explain why certain results are produced. Regulating a system that no one fully understands is a complex undertaking. But technical complexity cannot be used as a shield to evade criminal responsibility.
In Europe, the Digital Services Act, in force since 2022, requires large platforms to assess algorithmic risks, allow independent audits, and explain to users the logic of recommendations.
But after a long tug-of-war with Brussels lobbies, algorithm governance was left in the hands of the platforms: they decide whether the algorithms they design can be harmful, a vicious circle that is not yielding great results.
To overcome this contradiction, scholars are developing some concrete proposals.
The first provides that an independent public agency certify the safety of every algorithm before launch, especially if minors are involved, just as the Ministry of Health certifies the safety of a drug.
The second establishes that those who design an addictive product are civilly and criminally liable for the damage caused. In this case, the rule acts after the damage has been produced. The California jury recognized this principle judicially; the next step is to codify it into law.
This has been discussed for years, and the fact that Europe has not yet reached an acceptable result is mainly due to pressure exerted by lobbyists.
LIKE CIGARETTES
The fact that technology platforms are now more powerful than states has very important consequences.
The Commission has repeatedly admitted it has not proposed more radical solutions because platforms provided negative feedback during consultations.
Their first defensive barrier is that overly strict rules would slow innovation.
The second is more subtle. According to the platforms, if the design of algorithms is questioned, an alternative design must be suggested so that the business model can continue to stand.
They argue that a “safer” design of algorithms would be less effective: if infinite scroll were replaced with “finite” scroll, it would not be possible to keep the user on the site. How can an industrial strategy that deeply affects companies’ profitability be imposed by law?
This essentially echoes the objection used for decades by the tobacco industry which – while denying the harmfulness of cigarettes – argued that making certain changes to its products would reduce sales and profits.
It was necessary to prove that smoking causes cancer – and to reveal that executives of multinational companies were aware of this – to overcome these objections.
Until today, when criminal liability was discussed, it referred to tangible products: a defective car, a poorly designed airplane engine, an unsafe scaffold, and so on.
The California court has implicitly declared that a platform aware of the damage its algorithms cause to adolescents – as META’s internal emails show and as Mark Zuckerberg had to admit – and that does not intervene, is in a position similar to that of an engineer who installs non-compliant structures.
A widely cited article, published in 2025 by the Columbia Law Review (Sabriyya Pate, Platform Liability for Platform Manipulation), suggests introducing the crime of “negligence in platform design” that would make social media companies responsible for damages caused to users. But for now, these are legal exercises that have not turned into laws.
THE POWER OF LOBBIES
The European Commission itself admits that delays in legislating platform behavior are linked to lobby actions.
Between October 2021 and January 2022, the Commission launched a consultation on a “Directive on Liability for Artificial Intelligence” proposing a mechanism called “reversal of the burden of proof”: it provided that platforms would have to prove they did not harm users, rather than the opposite.
Consumer organizations and consulted academic institutions overwhelmingly approved the directive. The only opposition came from the platforms.
In a document filed on September 28, 2022, on EUR-Lex, it can be read that, regarding “the reversal of the burden of proof (…) companies provided negative feedback during consultations.” Result: in October 2025 the Commission withdrew the Directive due to “lack of agreement” with companies.
This example says a lot about where the balance of power between institutions and platforms lies. On the other hand, it is no surprise.
Political parties (both in the United States and Europe) have willingly freed themselves from the burden of fighting companies that now manage information and communication.
They have delegated the problem to an international institution (the Commission) without creating awareness among citizens about the damage digital monopolies are causing to our social, cultural, political life; even to our health.
It is no coincidence that guidance came from an institution independent of politics: a California jury. Let us hope it helps.
(Excerpt from Stefano Feltri’s Notes)
