AI CSAM Policy Gaps Remain Unclear

Single-source brief: AlgorithmWatch reports fast growth in AI-generated CSAM, while key enforcement details remain unclear.

What changed

AlgorithmWatch reports that offenders increasingly use generative AI to create child sexual abuse material, or CSAM.

The Internet Watch Foundation, or IWF, has tracked this activity since 2023.

AI-generated material made up about 1% of CSAM analyzed by IWF in 2025.

That share was small, but reported growth was sharp.

IWF identified more than 260 times as many AI-generated CSAM videos in 2025 than in 2024.

Its 2024 figure was already 130 times higher than its 2023 figure.

The IWF says 97% of victims are girls.

How AI tools are used

IWF staff cited by AlgorithmWatch say offenders often use open-source AI models.

Those models can be downloaded, changed, and trained to produce specific results.

IWF identified nearly 202 models linked to victims known from earlier abuse imagery.

It also received 17 reports involving free chatbots on mainstream websites.

In 2024, 40% of AI-generated CSAM reported to IWF was Category A material.

The IWF calls Category A its most severe category.

Policy stage and open questions

AlgorithmWatch reports that the European Union approved an AI Act amendment in May 2026.

The amendment bans AI systems from making such images without consent.

The supplied evidence does not give an effective date.

It does not explain penalties, enforcement, or which providers must comply.

AlgorithmWatch says it is doubtful that the amendment alone will curb the trend.