Meta internal memos surface in 2026 youth-harm trial

The trial centers on claims that Instagram and Facebook algorithms were tuned to maximize session length despite internal research flagging elevated suicide ideation rates among users aged 13-17. Plaintiffs introduced A/B test results showing a 23% lift in time spent after removal of friction features previously flagged by safety staff. Parallel documents from Philip Morris 1990s litigation show identical metric-over-harm prioritization language.

Data released under subpoena include engagement dashboards tracking “problematic use” cohorts that overlapped 68% with clinical depression flags. Tobacco precedent established that internal knowledge of harm, once proven, shifts liability from negligence to willful conduct. Meta’s surveillance architecture logged scroll velocity and dwell time at millisecond resolution, enabling granular targeting that mirrors nicotine delivery optimization.

Operational impact includes mandated algorithm audits and potential default settings that cap daily exposure. Prior settlements in tobacco cases required disclosure of all internal research; similar orders here would expose ranking code and real-time user-state classifiers. Regulatory filings already reference these logs as evidence of systemic design choices rather than user behavior.

Next phase hearings scheduled for September 2026 will test admissibility of engagement telemetry. If admitted, discovery expands to 2024-2025 model updates that retained the same objective functions.