A policy line becomes a credibility test.
Los Angeles, February 2026.
Mark Zuckerberg took the witness stand in a civil trial that accuses major social platforms of building products that hook children, and he reached for a defense that sounds simple but carries sharp implications. He reiterated that users under 13 are not allowed on Instagram, then argued that enforcement is difficult because a meaningful number of people misrepresent their age to access the services. The remark triggered immediate controversy because it shifts the center of gravity from design choices to user deception, a move that can read as pragmatic inside a product team and evasive inside a courtroom. What is being tested is not only whether the rule exists, but whether the company’s posture treats underage access as an edge case or as a predictable feature of the environment.
The case is being watched because it is forcing a public, sworn confrontation between corporate narratives and the lived experiences that plaintiffs say their children endured. The lawsuit centers on claims that Instagram and YouTube, among others, used engagement mechanics that exploited adolescent vulnerabilities, keeping young users online longer than they intended and amplifying harmful feedback loops. Some defendants in the broader ecosystem have reportedly reached settlements in related litigation, leaving this proceeding to function as a kind of proxy battle over responsibility and defect design. In that setting, the CEO’s credibility becomes part of the evidentiary terrain, because jurors are not only absorbing facts, they are judging whether the witness sounds like someone confronting harm or managing reputational exposure.
Zuckerberg’s age comment sits at the intersection of three constraints that rarely align cleanly: legal compliance, technical enforcement, and behavioral reality. The legal side is straightforward in theory, platforms say they bar children under 13, in part because of child privacy rules and because younger users raise higher risk. The technical side is harder, because age gates often depend on self reported data and verification systems can introduce friction that reduces sign ups, especially for older teenagers and young adults. The behavioral side is the most uncomfortable, because a system that is easy to bypass will be bypassed, and a system that is hard to bypass will be accused of surveillance or exclusion. In court, “people lie” can sound like a confession that the current approach relies on weak controls by design.
The plaintiffs’ narrative is built to make that weakness feel less like an engineering limitation and more like a business choice. Their argument is not simply that underage users appear, but that the product environment rewards time spent, and that the most vulnerable users are the most likely to be captured by that reward structure. If a platform’s revenue model correlates with attention, then every enforcement gap risks being interpreted as an incentive gap. That is why a statement about lying ages does more than describe a hurdle, it invites the question of how aggressively the company tried to solve it when solving it could slow growth. A jury does not need to understand machine learning to understand motive.
From Zuckerberg’s side, the testimony has emphasized policies, tooling, and improvements, including efforts to identify underage users and remove them when detected. He has framed the work as ongoing, with the subtext that the company wishes it had reached certain detection capabilities sooner. This positioning aims to portray the problem as real, but solvable, and as something the company has been steadily investing in rather than ignoring. Yet the courtroom dynamic punishes abstractions. When grieving parents or harmed users are part of the public frame, incremental improvement can sound like delayed urgency, especially if internal documents and prior warnings are introduced to argue that the risk was known years earlier.
The most volatile element is how the word addiction is being handled. Meta executives have drawn a line between clinical addiction and what they call problematic use, a distinction that is defensible in medical terms but politically fragile in plain language. Parents tend to hear “not clinically addictive” as a semantic shield, while companies present it as precision that prevents moral panic and bad science. The trial effectively turns that semantic difference into a moral test: if a product reliably produces compulsive patterns in children, does it matter whether the label meets a diagnostic threshold, or does responsibility attach at the level of foreseeable harm. In front of jurors, language choices can either signal honesty or signal corporate distancing.
The age question also exposes a governance problem that extends beyond one company. A child can lie about their age on almost any open platform, but the scale and design of each service determine how much that lie matters. If an ecosystem is built around virality, metrics, and social comparison, then underage presence is not merely a compliance issue, it is a risk multiplier. That is why this case is being tracked internationally, including by European and British outlets, as a bellwether for how courts may pressure platforms to redesign defaults rather than merely publish policies. Even if the trial’s scope is domestic, the reputational and regulatory aftershocks can travel.
What makes the moment unusually sensitive is that it connects two types of accountability that companies usually keep separate. Product accountability asks what the interface nudges users to do, how recommendations behave, and whether safeguards are frictionless or performative. Leadership accountability asks whether executives treated harms as core risk, or as a public relations variable to be managed around earnings, politics, and brand. When Zuckerberg says underage enforcement is hard because people lie, he is answering a product question, but he is also stepping into a leadership question. The courtroom is deciding which question it cares about more.
A platform does not need to lose a case to lose control of the narrative. If jurors, parents, and regulators walk away believing that age rules function more like disclaimers than barriers, then the next phase will be pressure for verification, tighter defaults for minors, or liability theories that treat certain design patterns as inherently unsafe for children. If, instead, the defense persuades that the company acted reasonably, improved detection, and cannot fully solve identity problems without unacceptable tradeoffs, then the result may be narrower, focused on edge cases rather than systemic redesign. Either way, the testimony reveals the central tension of the social internet: the business thrives on scale, but scale makes weak gates morally indefensible when the users behind those gates are children.
Contra la propaganda, memoria. / Against propaganda, memory.