K-Monitor has co-signed an open letter addressed to the European Commission, the European Parliament, and the Council of the European Union, calling on EU institutions to reject the Digital Omnibus package in its current form. The letter, signed by more than 40 civil society organisations from across Europe, expresses serious concerns about proposed amendments to the EU Artificial Intelligence Act (AI Act) — the 2024 regulation establishing a risk-based framework for AI oversight, including binding transparency, human oversight, and accountability requirements for high-risk AI systems used in areas such as employment, credit scoring, and criminal justice.

The Digital Omnibus package, published by the European Commission in November 2025, proposes a series of amendments that would substantially weaken these safeguards. Most notably, the application of key obligations for high-risk AI systems would be delayed until the European Commission adopts implementing decisions confirming the availability of harmonised standards, a process not expected to conclude before December 2027. Beyond the timeline, the package would broaden self-classification mechanisms, enabling providers to exempt their systems from high-risk obligations without independent verification, creating a structural incentive to avoid compliance even where systems materially affect fundamental rights. The proposal would further centralise enforcement authority within the EU AI Office at the expense of national supervisory authorities, reducing the democratic accountability and local institutional embeddedness that national bodies provide. Additionally, binding AI literacy training obligations for businesses and deployers would be replaced with non-enforceable encouragement measures, and incident reporting and logging requirements, critical tools for identifying discriminatory or erroneous algorithmic decision patterns, would be significantly reduced.

K-Monitor considers these changes collectively incompatible with the principle of meaningful public oversight of algorithmic decision-making. The proposed amendments would curtail the ability of citizens, civil society organisations, and public authorities to monitor and challenge consequential automated decisions, a matter of direct relevance to transparency and anti-corruption work. Negotiations between the Parliament, the Council, and the Commission are ongoing, and K-Monitor,  together with its co-signatories, urges EU institutions to preserve the core accountability guarantees of the original AI Act in any final text.