EU agrees tougher child protection rules against AI-generated abuse

The agreement between the European Parliament and the Council updates legislation first adopted in 2011, reflecting the growing role of digital technologies and AI in facilitating abuse.

Under the revised directive, designing, adapting or distributing AI systems intended to generate child sexual abuse material would become a criminal offence. The updated rules would also cover deepfake abuse material, livestreamed child sexual abuse, sexual extortion, and the possession or distribution of instructions on how to commit such crimes.

The agreement also strengthens rules on consent. It clarifies that consent must be given voluntarily, cannot be inferred from silence, lack of resistance or a previous relationship, and can be withdrawn at any time.

Grooming offences would be expanded to cover situations involving coercion, threats or deception, including cases where offenders falsely present themselves as peers of the child.

Victim protection would also be strengthened through access to healthcare, legal aid, helplines, accommodation support and compensation mechanisms. The agreement also extends limitation periods, recognising that many victims need years or decades before reporting abuse.

The revised directive still requires formal adoption by the European Parliament and the Council before entering into force.

Why does it matter?

The agreement shows how EU criminal law is being adapted to AI-enabled and online forms of child sexual abuse. Criminalising AI systems designed to generate abusive material is especially significant because it targets not only harmful content but also the tools used to produce it. The revised directive also strengthens victim support and prosecution timelines, addressing the reality that many survivors report abuse years after it occurred.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!