Report finds that online child sexual abuse material is increasing in sheer volume, growing more complex, and capitalizing on tech's blind spots
Aisha Counts / Protocol : Source: WeProtect Global Alliance .
Context & Ripple Effects
WeProtect Global Alliance's report lands more than a decade after the PROTECT our Children Act, which a 2019 New York Times investigation found tech firms and government had largely failed to enforce — so the finding that abuse material is growing in volume and complexity is a verdict on that enforcement gap, not a new problem. The pandemic made it worse: distribution spiked across major sites including Facebook and YouTube when moderation capacity was constrained during lockdowns.
What makes this report a turning point is the 'tech blind spots' framing. The related record shows where those blind spots lead: an analysis of millions of Discord and Telegram messages mapped organized grooming-and-extortion networks operating in hard-to-moderate channels, and by mid-2025 the Internet Watch Foundation was flagging its first wave of AI-generated abuse videos — a complexity problem this report anticipated.
First-order effects
- Platforms including Facebook and YouTube face immediate pressure to scale detection beyond what human moderation sustained during the pandemic surge, while WeProtect's findings hand regulators evidence that voluntary compliance under the PROTECT Act framework has not worked.
- Encrypted and semi-closed channels — the Discord and Telegram spaces where grooming networks were documented — become the focal point of the 'blind spots' argument, forcing those companies to defend their moderation posture publicly.
Second-order effects
- Detection vendors and cross-platform hash-sharing efforts gain leverage as buyers, because volume growth plus AI-generated content raises the cost of manual review faster than headcount can absorb.
- Regulators in the UK and EU get a concrete evidentiary base to move child safety from platform discretion toward mandated access controls and reporting duties, with the Internet Watch Foundation's escalating counts serving as the annual scorecard.
Third-order effects
- If generation tools keep outpacing detection, the field shifts from removing real recorded abuse to policing synthetic content at chokepoints — model outputs, upload pipelines, and age-gating — rather than takedowns after distribution.
- Sustained failure under a voluntary regime points toward structural change: statutory safety duties with independent auditing, making child-safety compliance a fixed cost of operating consumer platforms rather than a moderation choice.
The trend: Online child safety is moving from reactive takedowns of recorded material toward mandated, upstream controls as volume grows and generative tools create abuse content no archive has seen before.