The Internet Watch Foundation (IWF) has issued a stark warning about the increasing threat posed by AI-generated child sexual abuse images as the UK prepares to host a major global AI summit.
The IWF revealed that criminals are producing “astoundingly realistic” images of child sexual abuse using artificial intelligence technology. This surge in fake content risks normalising horrific abuse and diverting law enforcement resources away from real victims.
Between May 24 and June 30, the IWF investigated 29 web addresses flagged for potential AI-generated abuse images. Of these, seven were confirmed to contain such content, including some classified as severe, with victims as young as three years old.
Investigators also uncovered an online manual written by offenders that instructs others on how to train AI systems to create even more lifelike abusive images. While no real children are involved in the creation of these images, their existence disturbingly normalises abuse and complicates efforts to identify genuine cases.
Prime Minister Rishi Sunak has announced the upcoming global AI summit, aiming to foster international cooperation on AI-related risks. The IWF is calling for new legislation specifically targeting AI-generated child abuse content, warning that unchecked technological advances are creating a dangerous online environment.
Susie Hargreaves, CEO of the IWF, stated: “AI is becoming increasingly sophisticated. We need urgent, proactive measures. This threat must be a top priority at the global AI summit.”
The National Crime Agency (NCA) has confirmed that the threat is real and growing, although it remains limited in scale at present. The NCA reiterated that creating or possessing AI-generated child abuse images is a criminal offence under UK law.
The agency warned that such images normalise abuse and risk overwhelming law enforcement efforts, potentially delaying protection for actual child victims. The message from law enforcement and advocacy groups is clear: the darker capabilities of AI must be stopped before it is too late.
Originally published by UKNIP.
