Reports of AI-Generated Child Abuse Material Surge Over 8,400%, Research Shows
Reports of AI-driven child sexual abuse material rose from 4,700 in 2023 to over 400,000 in 2025, according to research from Childlight examining data submitted by tech firms.

Reports tracking the creation and dissemination of artificial intelligence-related child sexual abuse material escalated by more than 8,400% across a two-year span, according to research released by the Edinburgh-based Childlight – Global Child Safety Institute. Based on an analysis of disclosures submitted by technology firms to the US National Center for Missing and Exploited Children, recorded instances jumped from approximately 4,700 in 2023 to over 400,000 in 2025. This expanding volume includes synthetic media generated by so-called "nudify apps," which individuals leverage to extort minors for money or coerce them into providing further explicit content.
The Rapid Expansion of Synthetic Media
The institute's investigation focused on regions across North America, Latin America and the Caribbean, as well as East Asia and the Pacific, building upon earlier assessments that surveyed Western Europe. The trajectory revealed in Childlight's report aligns with corresponding findings from the Internet Watch Foundation (IWF), a UK-based non-governmental organization that tracks unlawful imagery internationally.
IWF reported that its investigators reviewed more photorealistic synthetic abuse material during the initial six months of 2026 than during the entirety of 2025. The organization assessed 6,310 AI-generated images that met the statutory definition of child sexual abuse, representing a 40% rise over the preceding year's tally of more than 4,500 images. Analysts observed that the prevalence of this material had already demonstrated marked expansion throughout 2025.
Widespread Harm and Systemic Pressures
Childlight calculates that over 300 million children worldwide experience at least one incident of technology-enabled sexual exploitation or abuse annually. Across a broader timeline, the institute estimates that approximately one in four individuals will endure some form of abuse prior to reaching 18 years of age.
The research also tracked offline abuse data alongside digital reporting to provide context on physical harm. Within North America, the study documented that nearly one in three girls and over one in ten boys have suffered in-person abuse during their youth, spanning offenses including sexual assault, rape, and "upskirting." Concurrently, administrative notices and formal reports regarding child sexual abuse material climbed by 63% in the United States between 2024 and 2025, compared to an 18% increase observed in Canada.
Prof Debi Fry, the global director of data at Childlight, characterized the findings as reflecting a "shocking" magnitude of exploitation. Fry observed that while enhanced identification protocols and heightened reporting practices by companies could account for a portion of the statistical spike, the sheer velocity of AI-driven changes presents a deep concern. She emphasized that governments and technology providers must urgently collaborate to understand these trends, intervene early, and expedite the removal of illicit media.
Paul Stanfield, Childlight's chief executive, urged state authorities to treat the crisis as a "preventable public health emergency." Stanfield noted that public entities frequently operate without comprehensive data regarding where harm occurs and who faces the highest risk, arguing that improved evidence gathering and coordinated global leadership are essential to protect youth.
What it means for developers
For software engineers and platform builders, these findings emphasize the acute regulatory and technical scrutiny directed at generative media tools. As automated systems become more capable of rendering convincing synthetic imagery, developers are under heightened pressure to implement rigorous guardrails, block prompt misuse, and cooperate with law enforcement and reporting agencies like the National Center for Missing and Exploited Children.
Teams integrating multimodal models or image-generation pipelines must construct proactive filtering layers to prevent malicious applications, including deepfake generation and nudify tooling. When testing and benchmarking safety filters across multiple foundational engines, developers can evaluate top AI models cheaply through one API at https://apixoai.online to determine which model architectures offer the most reliable safeguards against illicit prompts.
Beyond prompt sanitization, platform architects working with user-uploaded media face growing demands to deploy detection pipelines capable of distinguishing and flagging photorealistic AI abuse material rapidly. As data from organizations such as the IWF and Childlight demonstrates, monitoring infrastructure must scale quickly to handle the sharp increases in automated harmful content and support compliance with evolving global reporting requirements.
Source: Reports of AI-related child sexual abuse material rise 8,400% in two years, data shows — The Guardian AI. Written by the Apixo team from that report.
One key for Claude, GPT, GLM, DeepSeek and more. Pay per token with crypto.
Get your API key

