Arturo Bejar, a former safety engineer at Meta Platforms, has cast doubt on whether a landmark settlement between the technology company and a coalition of American states will genuinely protect teenagers from harm on its platforms. Speaking after eight jurors and a judge heard his testimony, Bejar suggested that the $18 billion deal announced by U.S. state attorneys general does not substantively remedy the problems he described during court proceedings — particularly Meta's systematic underreporting of teen harm, its algorithmic promotion of content damaging to adolescent self-esteem, and deficiencies in how the company investigates child exploitation.
California Attorney General Rob Bonta and his counterparts hailed the settlement as transformative, with Bonta declaring it would meaningfully reduce risks to young social media users. Yet Bejar's assessment presents a fundamentally different picture. He characterised the agreement as likely to entrench what he termed "safety theater," suggesting that while Instagram usage might decline modestly, the fundamental safety infrastructure of the platforms would remain unchanged. His scepticism stems from intimate knowledge of Meta's internal operations and the company's historical approach to similar interventions.
The settlement represents a rare moment of accountability for Meta, imposing financial penalties and requiring operational changes to Facebook and Instagram. Among the proposed measures are restrictions on how adolescents can engage with the platforms, enhanced parental oversight tools, and limits on daily usage. However, examination of Meta's own internal research reveals a troubling pattern: several remedies secured in the settlement had previously been evaluated by the company and rejected as insufficiently consequential to justify implementation.
A prime example concerns the hiding of like counts and reaction metrics on posts. This measure was designed to mitigate "negative social comparison" — the psychological harm teenagers experience when evaluating their content's performance against peers. Meta's internal testing, catalogued under the project codename "Project Daisy" dating to 2019, demonstrated that eliminating visible engagement metrics produced only marginal improvements in user wellbeing. Internal Facebook presentations to Mark Zuckerberg in 2020 revealed that researchers observed essentially no movement in overall wellbeing indicators when likes were hidden. Business impact modelling suggested that implementing this feature would reduce daily active users by approximately 0.09 percent — a negligible commercial consequence. Ultimately, Meta allowed users to voluntarily hide likes as an opt-in setting rather than adopting it as a default.
Bejar's own research at Meta had uncovered that teenagers experienced adverse effects on the platforms at rates substantially exceeding the company's publicly disclosed prevalence estimates. This disparity between internal findings and external communications formed a pillar of the states' legal case. Yet the settlement document addresses these allegations only peripherally, focusing instead on expanded parental controls and enhanced detection systems for underage users falsely claiming adult status. This represents a significant gap between the allegations prosecutors advanced and the remedies the settlement actually mandates.
Meta's algorithms have long been flagged by the company's own researchers as problematic for adolescent mental health. Internal studies identified the systems' propensity to inundate teenagers with intensive streams of fitness and beauty content — material that disproportionately affects those already struggling with self-image concerns. Additionally, the platform's design architecture intentionally extends user engagement beyond what individuals would choose independently, effectively engineering addictive patterns of behaviour. These design choices, which prioritise engagement metrics and advertising revenue, have persisted despite internal acknowledgment of their psychological toll on young users.
The challenge of verifying user age further undermines confidence in enforcement mechanisms. Meta relies fundamentally on user self-reporting for age verification, creating substantial loopholes. This structural weakness became apparent in Australia, where the government implemented a total prohibition on social media access for users under sixteen. Despite the legal ban, regulators found that approximately eight in ten young teenagers maintained active accounts on platforms including Meta's services, prompting authorities to contemplate additional enforcement measures.
Responses from clinical researchers and child mental health professionals to the settlement have remained divided. Dr. Jane Conron, a clinical psychologist at Northwestern University's Feinberg School of Medicine, noted that the requirement for Meta to provide a non-algorithmically curated feed to teenagers operates only on an opt-in basis — a provision she predicted would attract minimal uptake from users accustomed to algorithmic curation. Conversely, Conron acknowledged potential utility in the daily usage caps imposed on Instagram and Facebook for adolescent users. She observed that some of her young patients experience severe emotional distress when attempting to disengage from the applications, manifesting in tears and anxiety when parents enforce restrictions. Built-in technical limits could provide neutral enforcement without triggering the interpersonal conflict that currently characterises family negotiations over screen time.
Conron perceived incremental value in the settlement's very existence, despite its failure to require Meta's admission of wrongdoing or acknowledgment that its products inflicted documented harms. She suggested that the legal action and resulting agreement might catalyse broader societal recognition of social media risks, potentially reshaping how teenagers and parents discuss these platforms. The resolution signals to public consciousness that the issue warrants serious regulatory attention, even if the specific remedies remain modest in scope.
For Malaysian and Southeast Asian observers, this settlement carries particular significance. The region has witnessed accelerating teenage adoption of Meta's platforms, with Facebook and Instagram serving as primary social connection tools for adolescents across Malaysia, Indonesia, Thailand, and beyond. The regulatory architecture emerging from American litigation may influence how local governments approach similar concerns. Additionally, as Meta operates these platforms globally with largely uniform algorithms and features, protections implemented in North America theoretically extend to Asian users, though regional content moderation practices remain inconsistent. The debate surrounding this settlement reflects fundamental questions about whether technology companies will voluntarily prioritise youth mental health over engagement-driven revenue models, or whether only mandated regulatory intervention can produce meaningful change.
