The European Council is scheduled to formally adopt the Digital Omnibus on AI today, June 29, 2026, completing the legislative process that began with a provisional political agreement on May 7 and European Parliament approval on June 16. Once the Council vote is recorded and the amended regulation published in the EU Official Journal, the new compliance framework takes effect three days later.
For HR and people-function leaders, the single most consequential change is the deferral of Annex III high-risk AI compliance requirements. Systems facing an August 2, 2026 deadline now have until December 2, 2027 to achieve full compliance. That is a 16-month extension on a standard that covers virtually every AI-assisted employment decision: recruitment and candidate screening, promotion and termination decisions, performance evaluation tools, worker monitoring systems, and task allocation platforms.
What the extension does not change
The deferral applies specifically to Annex III standalone systems. It does not touch Article 50 of the original AI Act, which requires providers and deployers to disclose when users interact with an AI system. That transparency obligation activates on August 2, 2026, as originally scheduled.
Advertisement
300 × 250
The extension also does not affect AI systems embedded in regulated products under Annex I, such as medical devices or safety-critical machinery. Those systems now face a compliance deadline of August 2, 2028.
The scope of high-risk employment AI
Under Annex III, AI systems used in employment contexts are classified as high-risk when they influence decisions that affect individuals’ livelihoods. The categories are broad: any AI that screens job applications, selects interview candidates, evaluates performance, allocates tasks or shift schedules, monitors workers, or informs decisions on promotion, pay, or termination falls within scope.
For most large employers using AI-augmented applicant tracking systems, performance management software with machine learning scoring, or workforce analytics tools that feed into manager recommendations, the Omnibus extension is directly relevant to technology they are already deploying or actively evaluating.
AI literacy obligations
The Digital Omnibus also modifies Article 4, which covers AI literacy requirements for organizations deploying AI systems. The original language obligated providers and deployers to ensure staff achieve a defined level of AI competence. The amended text softens this to a requirement to support the development of AI literacy among staff, removing the guarantee of specific attainment and giving HR and L&D teams more flexibility in how they design training programs.
That change matters practically: the harder-edged original formulation implied that companies could be liable if an employee demonstrated insufficient AI literacy. The amended version shifts the obligation toward good-faith effort and process rather than demonstrated competency benchmarks.
For HR leaders who have been building AI training curricula in anticipation of the August 2026 deadline, the softened language does not make those programs less necessary. It means that the standard against which those programs will be judged in enforcement actions is now framed around intent and process rather than specific outcomes.
A new prohibition with an earlier deadline
One provision in the Digital Omnibus carries a deadline that arrives sooner than December 2027. The regulation adds a new prohibition under Article 5 on AI systems that generate or manipulate non-consensual intimate imagery and child sexual abuse material. That prohibition takes effect December 2, 2026.
For HR and legal teams assessing AI vendor contracts and acceptable-use policies, this new prohibition adds a category of explicitly banned capability. Any generative AI tool deployed in a workplace context that is capable of producing non-consensual imagery, even if that is not its intended use, needs to be evaluated for compliance with this ban before the December 2026 deadline.
What the extension period should look like
Compliance counsel generally caution that the extension should not be read as a pause. Documentation obligations, conformity assessment processes, and human oversight requirements do not become less complex by being deferred. Organizations that treat December 2027 as the new starting gun rather than the new finish line will face the same readiness scramble that the original deadline was creating.
HR technology teams that have already made progress on AI impact assessments, data governance frameworks, and employee notification processes are in a stronger position regardless of when formal enforcement begins. The Omnibus extension may relieve pressure on the August 2026 readiness date, but the competencies built in preparing for it have operational value independent of the compliance timeline.
For European employers with operations in the United States, the compliance picture is also shifting at the state level. California’s AI workforce compliance requirements are already imposing their own deadlines on HR systems, with obligations that run on a separate track from the EU framework.
Employers with AI-augmented hiring tools deployed in U.S. jurisdictions are simultaneously managing evolving judicial standards. A recent federal court ruling applied California employment law to AI screening vendors, setting a new precedent for vendor liability that runs in parallel with the EU regulatory track.
The bottom line
The Digital Omnibus on AI gives HR leaders breathing room on the high-risk compliance timeline. It does not reduce the scope of what compliance ultimately requires. The categories of AI that count as high-risk employment tools remain unchanged. The obligations that will eventually apply to those systems remain unchanged. What changed is the deadline and the AI literacy standard, not the destination.
Organizations that use the extension period to build governance structures, assess vendor compliance readiness, and train HR and legal staff will be far better positioned in December 2027 than organizations that defer that work along with the deadline.
Source: Council of the European Union
