We submitted a public comment to the European Commission on its draft guidelines for classifying high‑risk AI systems under Article 6 of the AI Act, focusing on Annex III point 8(b) on systems that can influence elections and voting behaviour. While welcoming the initiative, we warn that a narrow focus on systems explicitly “intended” for electoral use and on “direct exposure” to outputs risks overlooking powerful general‑purpose and backend tools that already shape voters’ choices and electoral infrastructure.
Our submission argues for a capability‑ and risk‑based interpretation of point 8(b), so that deepfakes, psychographic microtargeting and hidden voter‑suppression tools are reliably captured as high‑risk.
Read our full submission here.