The Digital Omnibus on AI: How Does It Impact Employment Relations?

On 24 July 2026, Regulation (EU) 2026/1744 of the European Parliament and of the Council of 8 July 2026, known as the "Digital Omnibus on AI", was published in the Official Journal of the European Union. This Regulation amends Regulation (EU) 2024/1689 (the AI Act) with the aim of streamlining and simplifying the application of the harmonized rules on artificial intelligence.

Postponement of High-Risk Obligations

Among other matters, this Regulation has resulted in the postponement of a significant portion of the AI Act's provisions relating to high-risk AI systems. Specifically, the provisions of Chapter III, Sections 1, 2 and 3 of the AI Act concerning standalone high-risk AI systems (classified under Article 6(2) and Annex III), which were due to enter into force on 2 August 2026, will now apply from 2 December 2027. In turn, the provisions relating to high-risk AI systems embedded in products (classified under Article 6(1) and Annex I), originally scheduled to enter into force on 2 August 2027, will now enter into force on 2 August 2028. For a comprehensive overview of all the changes introduced by the Digital Omnibus on AI, please refer to the Legal Flash | Key Aspects of the AI Omnibus Regulation.

How does this development impact the employment field? With the publication of the final text of the Digital Omnibus on AI, the postponement by 16 months of the entry into force and enforceability of the rights and obligations that the AI Act imposed for the use of high-risk AI systems in employment is confirmed, with exceptions such as those noted below.

That is, the obligations set out in the AI Act intended to address, among other matters, transparency, human oversight, fundamental rights impact assessments, and prior information to be provided to employees' legal representatives and the workforce will not be enforceable until 2 December 2027 for standalone systems (such as those used for recruitment, personnel management, and performance evaluation), and until 2 August 2028 for systems embedded in products.

The Digital Omnibus on AI does not only postpone obligations but also introduces substantive amendments that are already in force or will be shortly.

Omnibus developments that are already applicable

• Modification of the AI literacy duty
  Article 4 of the AI Act imposed on all providers and deployers of AI systems the obligation to ensure the AI literacy of their staff. The Digital Omnibus on AI has amended this provision, softening the obligation: providers and deployers, with the support of the Commission and the Member States, must now "take measures to support the development of AI literacy" of their staff, without being required to guarantee a specific level of literacy for any particular individual. Recital 8 of the Digital Omnibus on AI justifies this amendment by noting that the experience shared by stakeholders showed that a solution imposing strict obligations to ensure a sufficient level of AI literacy was not suitable for all types of providers and deployers, and that such obligations created an additional compliance burden, particularly for small businesses. See the Cuatrecasas Guide on The AI Literacy Obligation in EU Companies.

• New prohibitions under Article 5
  The Digital Omnibus on AI has added new prohibitions to Article 5 relating to the generation of non-consensual intimate material and child sexual abuse material, which will be applicable from 2 December 2026.

Obligations that remain in force

In addition to the new developments introduced by the Omnibus, both the AI Act and other existing legislation already contain specific provisions on AI that have been enforceable for some time, such as:

• The duty to provide algorithmic information to employees’ legal representatives, in line with article 64.4.d) of the Workers’ Statute in Spain, and, in Portugal, individually to each of the affected employees, and to the works council and the trade union representatives, in line with articles 106.3.s), 424.1.j) and 466.1.d) of the Portuguese Labor Code. For more information, see the post The duty to provide algorithmic information does not include revealing the source code.
  The prohibition on automated individual decisions under article 22 of the General Data Protection Regulation, which prevents decisions being adopted exclusively based on automated processing – including those adopted through AI systems – with significant legal effects for the employees, unless one of the exceptions established in that regulation applies.

• The provisions introduced by a collective bargaining agreement.

Conclusion

Ultimately, the postponement resulting from the Digital Omnibus on AI provides additional time to prepare for a significant portion of the obligations introduced by the AI Act in relation to high-risk AI systems for employment. However, this additional time should not be interpreted as a pause, given that the obligations already in force – such as AI literacy, algorithmic information requirements, the prohibitions under Article 5, and the GDPR's prohibition on automated decision-making – remain applicable, whether fully or with nuances. Organizations that precisely address these already-applicable obligations before 2 December 2027 will be better positioned to meet the AI Act obligations enforceable from that date onward.

Finally, it is worth recalling that the provisions of the AI Act constitute in any event a minimum floor that may be raised by Member States and, of course, by collective bargaining. Accordingly, attention should be paid to the regulation within collective agreements, which are increasingly incorporating provisions regarding the use of AI in the workplace.

For more information, please do not hesitate to contact our specialists through the Knowledge and Innovation Area.