For too long, children exposed to sexual abuse have faced a legal framework that has struggled to keep pace with the digital world. After more than two years of intense negotiations, the European Parliament and the Council have reached agreement on the recast Child Sexual Abuse Directive.
As rapporteur, I led Parliament's work on this file. This is not simply a technical update of 15-year-old rules, but a recalibration of European law to address threats that did not exist when the original directive was drafted.
This update to criminal law harmonizes child sexual abuse offences and penalties across the EU, including for crimes committed online, while strengthening victims' rights and access to support.
Member states must now adapt their national criminal laws into line with the revised directive.
This reform complements the proposed Child Sexual Abuse Regulation, which would require online platforms and service providers to detect, report and remove child sexual abuse material. Negotiations on the proposal, however, are still ongoing.
What the recast directive changes
One of the directive's key achievements is a new criminal framework for artificially generated child sexual abuse material.
Until now, those who possess AI-generated child sexual abuse material or design AI tools to generate it have operated in a legal grey area in some member states.
That will change under the recast directive. Such material will be illegal across the EU, and anyone who designs or distributes an AI system intended for that purpose will face criminal liability in every member state.
This was not an easy provision to negotiate.
Parliament was determined to strike the right balance: the directive does not criminalize general-purpose or open AI technologies, nor does it restrict their legitimate use. Instead, it targets intent and design.
A system built or adapted for the explicit purpose of generating abuse material poses a danger that has nothing to do with legitimate innovation and everything to do with the deliberate creation of tools for exploitation. This distinction ensures the legislation protects both children and the legitimate AI sector.
An urgent response
I was struck during a visit to the Internet Watch Foundation by how quickly the AI threat has grown.
In 2025, the IWF recorded a 154% rise in reports of realistic AI-generated child sexual abuse imagery and saw such videos increase more than 260-fold in a single year.
Their message was clear: this is not the crude, obviously synthetic content many associate with generative AI. Instead, it is often indistinguishable from real abuse. It traumatizes children when their likeness is used, desensitizes offenders and often depicts the most severe forms of abuse.
That visit reinforced my conviction that legislators could not afford to let technology outpace the law a second time.
This provision sits alongside other long-overdue reforms: the criminalization of so-called instruction manuals for abuse, livestreamed abuse and synthetic material, as well as extended limitation periods that enable survivors of child rape to seek justice until they reach the age of 50.
None of this happened by accident.
It reflects sustained cooperation across the political spectrum in Parliament, constructive engagement with the Council and, most importantly, the tireless efforts of civil society organizations. Without their persistence, this ambitious text would never have been possible.
The law is only as strong as its implementation. Member states must now transpose these rules swiftly and fully so they can take effect. Children cannot wait, and neither should European institutions.
Sign up to The Parliament's weekly newsletter
Every Friday our editorial team goes behind the headlines to offer insight and analysis on the key stories driving the EU agenda. Subscribe for free here.