Skip to main content

Joint letter: the EU’s Code of Practice for General Purpose AI cannot abandon children

In a joint letter sent this week to Executive Vice-President Virkkunen, 5Rights Foundation and over 15 organisations and experts — including Avaaz Foundation, the Ada Lovelace Institute, and ARTICLE 19 — have urged the Commission not to sideline fundamental rights and children’s protections in the third draft of the EU’s Code of Practice for General Purpose AI.

The call follows alarm over recent changes to the Code that relegate protections against child sexual abuse material (CSAM), nonconsensual intimate imagery, and other systemic harms to a voluntary appendix. Campaigners say this shift not only weakens the Code but risks undermining the very principles laid out in the AI Act.

The Code of Practice, they argue, still holds the potential to reflect the spirit of the AI Act, setting a global benchmark for human-centric and trustworthy AI governance. By enshrining protections for fundamental rights, democracy, and transparency, it can ensure that AI strengthens, rather than erodes, the freedoms that define open societies. But with growing pressure to dilute the Code in the name of consensus, the letter warns that compromise must have limits — and that limit is the faithful implementation of the EU AI Act.