Meta faces a four-front reckoning over child safety, privacy and AI copyright

Regulators, creators and publishers are challenging how Meta designs its platforms, uses public content and trains its AI models.

Meta is facing widening scrutiny on several fronts as regulators, courts and publishers press the company on child safety, data use and copyright. In Europe, the European Commission has already opened formal proceedings over the protection of minors on Facebook and Instagram, saying the platforms’ age-check systems may be too weak and warning that their recommendation systems could encourage addictive behaviour in children. The Commission’s later findings, according to its own statements, also questioned whether Meta had properly assessed the impact of features such as infinite scroll, autoplay and push notifications on users’ physical and mental wellbeing.

Those concerns have sharpened the debate over how social media platforms are designed. The Commission has said Meta’s risk assessments did not sufficiently account for evidence on children’s vulnerability to online harm, and it has questioned whether the company’s mitigation tools, including time-management features and parental controls, are strong enough to offset engagement-driven design. Ursula von der Leyen has since said the Commission plans draft legislation in the autumn to tighten age-appropriate protections online, following recommendations from an independent expert panel that called for tougher limits on younger users’ access to platforms with AI chatbots and similar features.

Meta is also under pressure over an AI image-generation tool on Instagram that can draw on content from public profiles. According to the company’s disclosures, accounts belonging to users under 18 are excluded and adults with public profiles can opt out, but the feature has nonetheless raised fresh privacy concerns because it uses faces and content from visible posts without requiring direct notice to the account holder. At the same time, the company is defending itself in India, where two creators have asked the Delhi High Court to examine alleged abuse of Facebook’s copyright enforcement system. They say the platform’s editing tools can be used to create backdated posts that trigger Rights Manager complaints against legitimate creators, a claim Meta has yet to publicly address in detail.

The company is also fighting a major copyright case in the United States. Five large publishers and author Scott Turow have accused Meta in federal court in Manhattan of using millions of copyrighted books and articles to train its Llama models without permission, with the complaint alleging that the company relied on pirated material and removed copyright management information from some works. That lawsuit follows earlier litigation over Llama training data and comes after a federal judge in California dismissed a separate case brought by 13 authors, while making clear that the ruling did not amount to a finding that Meta’s conduct was lawful. Taken together, the disputes suggest that Meta is now confronting a broader reckoning over how its products are built, how its systems are policed and how aggressively its AI ambitions are being financed by existing content.