Open letter urges EU lawmakers to retain AI Act transparency safeguard

More than 50 organisations and individual experts have signed an open letter calling on EU legislators to reject proposed amendments in the AI Omnibus that would remove a transparency requirement for high-risk AI systems.

Open letter urges EU lawmakers to retain AI Act transparency safeguard

A coalition of civil society organisations, advocacy groups, trade unions, and academic experts has issued an open letter to Members of the European Parliament, representatives of EU member states, Executive Vice-President Henna Virkkunen, and Commissioner Michael McGrath. The letter calls for the rejection of proposed amendments in the AI Omnibus that would delete Article 49(2) of the EU AI Act.

The signatories argue that Article 49(2) functions as a transparency safeguard for high-risk AI systems. Under Article 6(3) of the AI Act, providers of systems listed as high-risk in Annexe III may determine that their system does not pose a significant risk and exempt themselves from the obligations applicable to high-risk AI. Article 49(2) requires providers using this exemption to register their decision in a publicly accessible database.

According to the letter, removing this requirement would create a loophole allowing companies to opt out of high-risk obligations without public oversight. The signatories state that deletion of the safeguard would weaken enforcement, reduce transparency for market surveillance authorities, undermine harmonisation within the EU Single Market, and limit the ability of affected individuals and civil society to scrutinise exemptions.

The European Commission’s accompanying Staff Working Document estimates that removing the registration requirement would save companies an average of €100, reflecting approximately 2.5 hours of administrative work. The signatories argue that this cost reduction does not justify what they describe as a significant weakening of accountability and protection mechanisms.

The letter specifically calls on EU legislators to reject paragraphs 6, 14, and 32 of the AI Omnibus to restore the Article 49(2) transparency safeguard.

Signatories

  • Access Now
  • European Digital Rights (EDRi)
  • 5Rights Foundation
  • AI Forensics
  • AK Europa
  • AlgorithmWatch
  • Alternatif Bilisim
  • Amnesty Tech
  • ARTICLE 19
  • Asociația pentru Tehnologie și Internet
  • Association pour la taxation des transactions financières et pour l’action citoyenne (Attac)
  • BEUC – European Consumer Organisation
  • Bits of Freedom
  • Centre for Democracy and Technology Europe
  • Coalition for Independent Technology Research
  • Corporate Europe Observatory (CEO)
  • Danes je nov dan, Inštitut za druga vprašanja
  • Defend Democracy
  • Democratic Society
  • Deutsche Vereinigung für Datenschutz e.V. (DVD)
  • Die Bürokratiemonster
  • Digitalcourage e. V.
  • Digitale Gesellschaft (Germany)
  • Digitale Gesellschaft (Schweiz)
  • Electronic Frontier Norway
  • epicenter.works
  • European Center for Not-for-Profit Law (ECNL)
  • European Civic Forum
  • European Council of Autistic People
  • European Disability Forum (EDF)
  • European Environmental Bureau
  • European Network Against Arms trade (ENAAT)
  • European Public Service Union (EPSU)
  • Fix the Status Quo
  • Gong
  • Homo Digitalis
  • Human Development Research Initiative (HDRI)
  • Investor Alliance for Human Rights
  • Irish Council for Civil Liberties
  • IT-Pol Denmark
  • La Quadrature du Net
  • Lafede – justícia global
  • Likestillings- og diskrimineringsombudet / Equality and Anti-Discrimination Ombud (Norway)
  • Metamorphosis Foundation for Internet and Society
  • Panoptykon Foundation
  • People vs Big Tech
  • Politiscope
  • Stichting Health Action International (HAI)
  • The Good Lobby
  • Weaving Liberation
  • Wikimedia Deutschland e.V.
  • younion _ Die Daseinsgewerkschaft
  • Dr. Gianclaudio Malgieri, eLaw Leiden University
  • Dr Abeba Birhane, AI Accountability Lab (AIAL), Trinity College Dublin
  • Dr. Harshvardhan Pandit, AI Accountability Lab (AIAL), Trinity College Dublin
  • Maribeth Rauh, AI Accountability Lab (AIAL), Trinity College Dublin
  • Dr. Zeerak Talat, University of Edinburgh
  • LK Seiling, Weizenbaum Institut / DSA40 Data Access Collaboratory
  • Dr. Laura Caroli
  • Dr. Aida Ponce Del Castillo
Go to Top