The emergence of Generative AI tools has introduced new methods for creating Child Sexual Abuse Material (CSAM). These technologies have enabled new forms of harm against children, while significantly complicating the tasks of victim identification and safeguarding.

Law enforcement is now faced with unprecedented challenges. Every image now prompts the difficult questions: Is this image AI-generated? Is it a real photograph that has been edited with AI? Is a real child being depicted in this image? Is there a child who needs our help? The realism of AI images has advanced to the point that these images are often visually indistinguishable from real photographs, yet many investigators are often left to rely on just visual judgement alone. If the only AI images you see are really obvious (like extra limbs or distorted faces), it is worth considering what might be being missed. CameraForensics Ember provides an extra pair of eyes, supporting law enforcement by detecting AI content accurately and at scale.

This talk will outline the key ways Generative AI is being used to produce CSAM, explain the implications for child safety, and highlight the complexities this creates for those working to detect and respond to abuse. We will demonstrate how CameraForensics is working to support investigators in tackling this evolving threat, through tools designed to detect, localise, and examine AI-generated content.