Resolver, a Kroll business and global leader in trust and safety intelligence, announced the launch of its Unknown CSAM Detection Service on July 17, 2025, in Leeds, United Kingdom. Powered by the Roke Vigil AI CAID Classifier, this service enables social and technology platforms to automatically identify and classify previously unseen and AI-generated child sexual abuse material (CSAM). By addressing the surge in AI-driven CSAM, the service sets a new standard for online child safety, offering high-precision triage, enforcement, and safeguarding at scale.
Resolver launches Unknown CSAM Detection Service to identify unseen and AI-generated CSAM.
Powered by Roke Vigil AI CAID Classifier, trained on the UK’s Child Abuse Image Database (CAID).
Categorizes CSAM by severity (A, B, C, or indicative) for effective triage.
Previously exclusive to law enforcement, now available to platforms via Resolver.
Addresses 2025 surge in CSAM reports, including AI-generated content.
Supports platforms with managed API and deployment for regulatory compliance.
The Internet Watch Foundation’s May 2025 report highlighted a record number of CSAM reports, with a significant rise in AI-generated material, underscoring the evolving threat landscape. In June 2025, the UK’s Ofcom recommended proactive technologies to detect unknown CSAM, signaling a shift from relying solely on known content detection. Resolver’s service responds to this challenge by identifying both known and previously unseen CSAM, including AI-generated imagery, which predators increasingly exploit. This capability ensures platforms can act swiftly to protect users and meet regulatory expectations.
The service leverages the Roke Vigil AI CAID Classifier, a machine learning solution trained on the UK Government’s Child Abuse Image Database (CAID) under strict Home Office oversight. Unlike traditional hash-matching, which only detects known CSAM, this classifier identifies new, first-generation, and AI-generated material with high accuracy. It categorizes imagery by severity (categories A, B, C, or indicative CSAM), enabling faster triage and routing high-risk content to specialized teams. Previously reserved for law enforcement to scan seized devices, this technology is now accessible to platforms through Resolver’s managed service.
The classifier’s effectiveness stems from two key advantages: access to CAID, the world’s most curated CSAM dataset, and rigorous verification processes. Each image in CAID is vetted by three specialist officers to meet evidential standards and categorized by abuse severity. Routine retraining ensures the model adapts to evolving abuse patterns, maintaining accuracy against emerging threats like AI-generated CSAM. This robust approach reduces false positives and enhances the precision of detection, critical for safeguarding children online.
Resolver’s service, delivered as a cloud-based solution with a managed API, integrates seamlessly into platform workflows, reducing the burden on Trust & Safety teams. “The Roke Vigil AI CAID Classifier sets a new standard for automated, scalable detection of CSAM,” said George Vlasto, Head of Trust & Safety Division at Resolver. The service includes deployment support and training to protect staff from traumatic content exposure, enabling rapid response and compliance with regulations like the UK’s Online Safety Bill. “This launch is another commitment from Resolver to keep children safe,” said Kam Rawal, President of Resolver.
Resolver’s Unknown CSAM Detection Service marks a significant advancement in combating online child exploitation. By leveraging AI to detect and categorize both known and unseen CSAM, the service empowers platforms to protect users, comply with regulations, and support law enforcement efforts. As AI-generated CSAM proliferates, Resolver’s solution, backed by the Roke Vigil AI CAID Classifier, offers a proactive, scalable approach to creating safer digital environments for children worldwide.
Resolver, a Kroll Business, a leader in risk intelligence solutions, safeguards $6.5 trillion in market cap across 1,000+ companies worldwide. Leveraging AI with deep human expertise, we help transform complex data into actionable insights, enabling businesses to mitigate risks, recover swiftly, and strengthen operations. Discover how at Resolver.com. See Risk. Build Resilience.
Roke is a world-class UK innovation company, combining science and engineering to apply new technologies to challenging problems. The Roke Vigil CAID AI classifier is a testament to Roke's commitment to leveraging advanced technology for societal good.
TechIntelPro empowers B2B technology decision-makers with the latest industry trends, deep insights, leadership perspectives and actionable intelligence to drive both organizational and professional milestones—and thrive on the tech pulse.