On 7 May 2026, the Council of the EU and the European Parliament reached a provisional political agreement on the Digital Omnibus on AI, a targeted simplification package proposed by the Commission in November 2025 (part of the broader Omnibus VII legislative package).
Key changes to Article 4 (AI Literacy):
- Original obligation: providers and deployers of AI systems must take measures to ensure a “sufficient level of AI literacy” of their staff and persons operating AI on their behalf — a binding requirement on organisations
- Amended obligation: the direct obligation on providers and deployers is removed; instead, the Commission and Member States are required to encourage providers and deployers to take literacy measures (e.g. training opportunities, informational resources, good-practices exchange)
- Parliament had sought to retain a mandatory obligation but lower the standard from “sufficient level” to “adequate understanding”; the provisional agreement shifts responsibility from organisations to public authorities
- Enforcement timeline changes: high-risk AI system deadline for Annex III (standalone systems, including employment AI) delayed from 2 August 2026 to 2 December 2027; Annex I (high-risk embedded in regulated products) delayed further to 2 August 2028
- A new prohibited practice covering AI-generated non-consensual sexual/intimate content applies from 2 December 2026
- Formal adoption required before 2 August 2026; not yet formally adopted as of 25 May 2026
Score relevance:
- Dim 6 (employers legally required to train employees): weakened from the original binding Article 4 obligation to an encouragement framework — score may need to move from 1 toward 0 once formally adopted
- Dim 11 (civil servants required to receive AI training): similarly weakened; encouragement replaces mandate — score may move from 1 toward 0
- Dim 12 (enforcement mechanism for AI literacy): the previously identified penalty (EUR 7.5M/1.5%) was tied to the binding obligation; softening the obligation reduces enforceability — current score of 1 may be revisited downward once formally adopted
- High-risk employment AI enforcement (supporting dim 9) is delayed to December 2027; dim 9 score of 2 remains valid as the legal framework is preserved, only the deadline shifts