Investigation Tool Adds Deepfake Detection to CSAM Work
The integration is announced, not shipped, and targets AI-generated abuse material that UNICEF, ECPAT and INTERPOL estimate reached at least 1.2 million children.
Updated on Nov. 2, 2026 in Artificial Intelligence

Amsterdam-based media forensics startup Cyberette has partnered with forensic software provider bizTops to integrate its deepfake detection platform into LUMIN, bizTops' visual intelligence system for investigating child sexual abuse material. The partnership was announced at the 42nd meeting of the INTERPOL Specialist Group on Crimes Against Children in Lyon, France. The integration is announced, not yet shipped.
Why it matters
Hotlines and law enforcement agencies use LUMIN to analyze large collections of images and video, and AI-generated material adds a layer of complexity to these investigations. Pairing manipulation detection with a CSAM classifier from Rigr AI targets the triage problem directly.
The estimate of 1.2 million affected children comes from a joint UNICEF, ECPAT International and INTERPOL study covering 11 countries over one year. The market forecasts (6 billion checks in 2026, 9.9 billion in 2027) are projections, not measured volumes, and the source gives no performance benchmark for Cyberette's detection accuracy.
The players
Cyberette
Amsterdam-based media forensics and deepfake detection startup whose platform flags suspected manipulation in images, video and audio.
bizTops
Forensic software provider that develops LUMIN, its visual intelligence system for investigating child sexual abuse material.
Rigr AI
Provider of the CSAM classifier that will be incorporated into LUMIN alongside Cyberette's detection technology.
INTERPOL
International police organization whose Specialist Group on Crimes Against Children hosted the 42nd meeting where the partnership was announced.
The details
Cyberette's platform looks for signs of suspected manipulation in images, video and audio, which is the core task of deepfake detection. Those signals will be combined inside LUMIN with a CSAM classifier from Rigr AI, a model trained to flag child sexual abuse material, and a pre-configured Universal Classification Schema, a shared labeling framework that sorts findings into standard categories. The result is a single workflow in which an investigator can screen a large batch of files for both synthetic manipulation and known abusive content.
Timeline
2026 — forecast of more than 6 billion deepfake detection checks globally.
2027 — forecast of 9.9 billion checks and nearly $5 billion in revenue.
The Tech Race
The integration places Cyberette's manipulation detection alongside Rigr AI's CSAM classifier inside a single investigative platform, a combination that the source presents as a response to the scale of the problem documented by UNICEF, ECPAT International and INTERPOL. The source does not name competing products, so the wider competitive landscape cannot be characterized from this announcement.
The integration is announced rather than shipped, and the source gives no availability date or pricing for agencies or hotlines. Practically, it changes how investigators are expected to triage large image and video collections once LUMIN incorporates the new detection and classification layers.
The takeaway
The partnership signals that AI-generated abuse material is being folded into the standard tools investigators already use, rather than handled as a separate problem. Watch for the first named agency or hotline deployment of the integrated LUMIN workflow, which will show whether the combined detection and classification approach reaches operational use.
Further reading
For broader context on how these systems are built and tested, see our coverage in the Artificial Intelligence section.
Source note: This article includes information reported by Biometric Update.


