🔗 Share this article

Tech firms and child safety organizations will receive permission to assess whether artificial intelligence systems can produce child abuse images under recently introduced UK legislation.

Significant Increase in AI-Generated Harmful Material

The announcement coincided with findings from a safety watchdog showing that reports of AI-generated CSAM have increased dramatically in the last twelve months, growing from 199 in 2024 to 426 in 2025.

New Legal Structure

Under the amendments, the government will permit approved AI developers and child protection organizations to examine AI systems – the foundational technology for chatbots and image generators – and ensure they have adequate protective measures to prevent them from creating images of child exploitation.

"Fundamentally about stopping abuse before it happens," stated the minister for AI and online safety, adding: "Experts, under strict conditions, can now identify the risk in AI systems early."

Addressing Regulatory Challenges

The amendments have been introduced because it is against the law to create and own CSAM, meaning that AI creators and other parties cannot create such content as part of a testing regime. Until now, authorities had to wait until AI-generated CSAM was published online before addressing it.

This legislation is designed to preventing that issue by enabling to halt the production of those materials at source.

Legal Structure

The changes are being added by the authorities as modifications to the criminal justice legislation, which is also implementing a ban on owning, creating or sharing AI models designed to generate child sexual abuse material.

Real-World Consequences

This recently, the minister visited the London base of a children's helpline and listened to a mock-up conversation to counsellors featuring a report of AI-based exploitation. The call depicted a adolescent requesting help after facing extortion using a sexualised deepfake of himself, constructed using AI.

"When I learn about young people experiencing blackmail online, it is a cause of extreme anger in me and rightful concern amongst parents," he stated.

Concerning Data

A prominent internet monitoring organization stated that cases of AI-generated abuse material – such as online pages that may contain multiple images – had significantly increased so far this year.

Instances of the most severe material – the most serious form of abuse – increased from 2,621 images or videos to 3,086.

• Female children were predominantly targeted, making up 94% of prohibited AI depictions in 2025
• Portrayals of infants to two-year-olds increased from five in 2024 to 92 in 2025

Sector Reaction

The legislative amendment could "constitute a vital step to ensure AI products are safe before they are released," stated the head of the online safety organization.

"Artificial intelligence systems have made it so survivors can be targeted repeatedly with just a few clicks, providing offenders the ability to create potentially endless quantities of sophisticated, photorealistic child sexual abuse material," she continued. "Material which additionally commodifies survivors' suffering, and renders young people, particularly female children, less safe both online and offline."

Counseling Interaction Information

The children's helpline also released details of counselling interactions where AI has been mentioned. AI-related harms discussed in the sessions comprise:

• Using AI to evaluate weight, physique and appearance
• AI assistants discouraging children from talking to trusted adults about abuse
• Being bullied online with AI-generated content
• Online blackmail using AI-faked pictures

Between April and September this year, Childline delivered 367 counselling interactions where AI, conversational AI and associated topics were discussed, four times as many as in the same period last year.

Half of the references of AI in the 2025 interactions were connected with psychological wellbeing and wellbeing, encompassing using chatbots for assistance and AI therapeutic applications.