Meta Platforms has reached an $18 billion settlement with a coalition of United States states over child safety concerns, but the agreement has drawn sharp criticism from Arturo Bejar, the former safety engineer whose damaging testimony helped build the case against the social media giant. Bejar argues that the settlement package, announced following what California Attorney General Rob Bonta called a breakthrough for protecting children online, fails to meaningfully address the substantive harms he was called to describe in court regarding how the company knowingly endangers teenagers on its platforms.
During his testimony before a judge and jury, Bejar laid out a troubling picture of Meta's practices. The company systematically underreported the frequency with which teenagers experienced harm across its platforms, he stated. More insidiously, Meta deliberately steered young users toward content designed to make them feel inadequate about themselves, while simultaneously failing to properly investigate and act upon reports of child predation. These allegations struck at the heart of whether Meta prioritised profit maximisation over genuine child safety protections, a question that has animated regulatory scrutiny worldwide, including across Southeast Asia where regulators increasingly scrutinise how global platforms handle minors.
Bejar's post-settlement critique carries particular weight given his insider status and central role in the legal proceedings. He characterised the agreement as likely to "codify much of the 'safety theatre' that Meta has been doing in recent years," suggesting that while Instagram usage among teenagers might marginally decline, the platform would remain fundamentally unsafe. This distinction matters enormously: the settlement focuses on reducing how much time young people spend on Meta's services, but not on whether those services, when used, actively harm their mental wellbeing. For Malaysian policymakers and parents watching how global authorities approach tech regulation, Bejar's scepticism underscores a critical gap between visible compliance measures and substantive risk mitigation.
The settlement's centrepiece remedies reveal the limitations of what US attorneys general secured through litigation. Meta has agreed to hide likes and reaction counts on posts, ostensibly to reduce harmful social comparison among teenagers who obsessively track their content's performance against peers. Yet internal Meta documents, revealed through legal discovery, demonstrate that the company explored this exact measure years earlier under the codename "Project Daisy." Meta's own researchers concluded the change would have at most modest impact on user wellbeing. Crucially, employee slide presentations to then-Chief Executive Mark Zuckerberg acknowledged that hiding likes would barely dent daily user counts, with projections suggesting only a 0.09 percent decline. Meta ultimately confined the feature to an opt-in setting, essentially rendering it ineffective since most teenagers never activate it.
This pattern of Meta considering and rejecting safety measures because they proved "relatively inconsequential" both for user wellbeing and corporate bottom lines reveals a troubling corporate calculus. The company's internal researchers had flagged that its algorithms disproportionately recommend fitness and beauty content to teenagers already struggling with self-esteem issues, thereby amplifying existing psychological vulnerabilities. Similarly, Meta engineers had documented that the platform's feature design deliberately encourages excessive scrolling beyond what users would freely choose. These algorithmic harms formed central planks of the states' case against Meta, yet the settlement addresses them only peripherally, focusing instead on parental controls, usage time limits, and tools to detect age misrepresentation.
Particularly troubling is Meta's ability to enforce age restrictions given that the platform relies on users to self-report their age. Australia's experience provides an cautionary example of how tech companies circumvent regulation. When Australia's government banned social media use by teenagers under 16, authorities subsequently discovered that approximately 80 percent of underage teens remained active on these platforms, prompting regulators to contemplate more aggressive enforcement mechanisms. For Malaysia and other regional governments considering similar age restrictions, the Australian precedent suggests that without rigorous identity verification and enforcement infrastructure, legislative solutions may prove largely theatrical.
The settlement's requirement that Meta cap daily usage hours for teenagers on Instagram and Facebook has attracted more cautious optimism from mental health specialists. Dr Jane Conron, a clinical psychologist at Northwestern University's Feinberg School of Medicine, noted that some of her adolescent patients experience such acute emotional distress when separated from these platforms that they cry when parents attempt restricting access. A built-in usage limit, she suggested, might provide meaningful relief by removing the friction between teen desire and parental enforcement. Such clinical observations highlight how Meta's design deliberately exploits psychological vulnerabilities to maximise engagement, suggesting that reducing exposure time may confer genuine health benefits beyond what algorithmic modifications could achieve.
Conron also flagged a significant limitation in the settlement: Meta must merely offer teenagers a non-algorithmically curated social media feed on an opt-in basis. She predicted this feature would go largely unused, as teenagers inherently resist choosing their own content curation rather than accepting whatever Meta's engagement-maximising algorithms deliver. This observation exposes a fundamental challenge in regulating platform behaviour through nudging rather than mandate: voluntary features designed to protect users typically fail because they demand active choice from populations psychologically conditioned to accept default settings. The settlement's reliance on opt-in protections thus undermines its stated intent to meaningfully improve teenage mental health outcomes.
Notably, the settlement structure avoids requiring Meta to acknowledge fault or accept liability for designing products intended to harm children. For corporate defendants, this represents a considerable advantage: Meta pays a substantial financial penalty while maintaining that its products operate as intended and are not inherently dangerous. This framing becomes crucial when considering how the company will approach future design decisions. Without an admission that its platforms cause measurable psychological harm, Meta faces no internal pressure to fundamentally reimagine its business model, which depends on maximising user engagement through algorithmic content curation and deliberate habit-formation features.
Regional implications for Southeast Asia merit consideration. Malaysia, alongside Singapore, Indonesia, and other ASEAN nations, has increasingly questioned whether international social media platforms adequately protect young users from algorithmic manipulation and mental health harms. The US settlement provides a data point about what regulatory action can realistically achieve when confronting platforms with extraordinary financial and legal resources. If $18 billion in penalties and nominal operational constraints satisfy US attorneys general, smaller developing economies may struggle to extract more meaningful commitments through their own regulatory efforts.
Some mental health experts, however, detected a silver lining in the settlement's mere existence. Conron suggested that the case and settlement itself might elevate public awareness about the genuine dangers platforms pose to adolescent mental health, thereby shifting conversations between teenagers and parents from vague concern to informed discussion grounded in documented harms. This broader cultural impact could prove more consequential than any specific remedial measure, as parent-teen dialogue informed by credible evidence of algorithmic manipulation might naturally constrain platform usage more effectively than any corporate-imposed restriction.
Yet Bejar's fundamental critique endures: the settlement amounts to managing a deeply problematic situation rather than resolving it. Meta's core business model—harvesting teenage attention and exploiting psychological vulnerabilities to maximise engagement and advertising revenue—remains intact and untouched by the agreement. The company will implement the required parental controls and usage limits, announce them as major safety achievements, and continue operating platforms whose fundamental design prioritises corporate profit over teenage mental wellbeing. For regulators, parents, and mental health professionals watching from Malaysia and across Southeast Asia, the settlement illustrates both the promise and profound limitations of litigation-based accountability when confronting technology giants whose business models structurally oppose meaningful youth protection.
