humanoid4rent
Prihlásiť saZískať cenové ponuky

Digital Omnibus and the AI Act: what was agreed, and what it means for robot operators

The AI Omnibus is now law. We explain the final text: later high-risk dates, a new place for machinery in the AI Act, softer AI literacy wording, and what is still unresolved.

Autor: redakcia humanoid4rent · zverejnené 14. novembra 2026

Táto stránka zatiaľ nie je preložená
Čítate anglický text, pretože verzia v jazyku Slovenčina zatiaľ neexistuje. Nič nebolo vynechané: toto je celá stránka v jazyku, v ktorom bola napísaná.

The EU institutions reached political agreement on the Digital Omnibus on AI on 7 May 2026, and the amending act, Regulation (EU) 2026/1744, entered into force on 27 July 2026. It moves the high-risk deadlines to December 2027 and August 2028, shifts machinery into a different part of the AI Act, and softens the AI literacy duty. For robot operators, most day-to-day obligations stay the same for now.

This article covers the final text and what is still open. For the full timeline of dates, see our earlier explainer on which AI Act rules apply to robots and when. It is general information about EU law, not legal advice.

From proposal to Official Journal

The procedure moved quickly by EU standards. The European Parliament's Legislative Observatory file 2025/0359(COD) records these steps:

  1. 19 November 2025 – the Commission published its proposal, COM(2025) 836, as part of a wider simplification package.
  2. 26 March 2026 – Parliament adopted its first-reading position after a joint vote of its internal market (IMCO) and civil liberties (LIBE) committees on 18 March.
  3. 7 May 2026 – Parliament and Council negotiators reached political agreement in trilogue, according to the Commission's AI Act policy page.
  4. 16 June 2026 – Parliament approved the agreed text in plenary.
  5. 29 June 2026 – the Council adopted it.
  6. 8 July 2026 – the act was signed; it was published in the Official Journal on 24 July 2026.

The Commission's announcement that the AI Omnibus enters into force gives 27 July 2026 as the date it took effect across the EU. The Omnibus is therefore no longer a draft. The open questions now concern implementation, not negotiation.

The new high-risk dates

The headline change is timing. The final text sets two dates for the high-risk rules, which the Commission lists as follows:

  • 2 December 2027 – high-risk rules for systems used in the sensitive areas listed in Annex III, such as biometrics, critical infrastructure, education, employment, migration and border control.
  • 2 August 2028 – high-risk rules for AI built into products covered by EU product-safety law. The Commission's entry-into-force announcement names machinery, toys and lifts as examples.

Under the original AI Act these dates were August 2026 and August 2027, so robots built into regulated products gain one more year.

Not everything was postponed. The general application date of 2 August 2026, including the transparency rules covered in our Article 50 explainer, was kept. The Omnibus also added a new ban on AI that generates non-consensual intimate content or child sexual abuse material, which the Commission says applies from December 2026.

Machinery has moved within the AI Act

For robots, the most significant change is less visible than the dates. Annex I of the AI Act lists the EU product laws that can make an AI system high-risk. It has two parts. For laws in Section A, the AI Act's high-risk requirements apply directly. For laws in Section B, such as those on cars and aircraft, they apply in a more limited way.

The consolidated Annex I on the AI Act Service Desk now shows the old machinery entry in Section A as deleted. The Machinery Regulation (EU) 2023/1230 appears instead as a new item 21 in Section B.

The consequence is set out in the amended Article 2(2). For high-risk AI linked to Section B products, only the classification rule in Article 6(1), a real-world testing article (60a) and the final provisions apply directly. The sandbox articles apply only to the extent that the AI requirements have been integrated into the sectoral law itself.

In practical terms, a humanoid or robot dog whose safety functions rely on AI should eventually be checked against AI-related requirements through the machinery rules, not through a separate AI Act conformity track. The Machinery Regulation already contains requirements for machines with self-learning safety functions, as we describe in our article on the Machinery Regulation and self-learning robots.

A narrower meaning of "safety component"

The Omnibus also clarifies when AI in a product counts as a safety component, which is what triggers high-risk status under Article 6(1). The amended Article 6 adds three new paragraphs:

  • AI that serves only non-safety purposes, such as user assistance, performance optimisation, service efficiency, automation, convenience or quality control, is not a safety component.
  • That exclusion does not apply where a failure of the AI would endanger health and safety.
  • A product that needs third-party assessment only for risks other than health and safety, such as radio spectrum use, does not meet the third-party condition on that basis alone.

For event robots this distinction is useful. A conversational assistant or a dance routine on a humanoid is likely to fall in the non-safety category. The software that keeps the robot balanced, stops it near people or plans its path through a crowd is a different matter, because its failure could hurt someone.

AI literacy: a softer duty

Article 4, the AI literacy provision, has applied since February 2025 and was reworded. According to the consolidated Article 4, providers and deployers must now take measures to support the development of AI literacy among their staff and others who operate AI systems on their behalf. A new paragraph requires the Commission and Member States to support these efforts, particularly for SMEs.

The Commission describes this as simplifying the previous requirement for companies, with public authorities taking a stronger role. The duty still exists, though. A team running a robot at an event should still be briefed on what the system does and where it can fail. Our AI literacy explainer covers what that can look like.

Other changes in the final text

The Commission's announcement lists several further points:

  • some SME simplifications, including lighter technical documentation, now extend to small mid-cap companies;
  • regulatory sandboxes are expanded, including a new EU-level sandbox;
  • EU database registration is streamlined for systems that are exempted from high-risk status;
  • processing of special categories of personal data is permitted for detecting bias;
  • the AI Office receives wider oversight of general-purpose AI models and of AI systems on large online platforms.

Most of these matter to developers and manufacturers rather than to the companies that hire robots.

What is still open

The law is settled, but several practical questions are not.

  • How machinery will be handled in practice. The Commission's announcement still lists machinery under the August 2028 date, while Annex I now routes it through the Machinery Regulation. How the AI requirements are written into the machinery rules, and when, will decide what manufacturers must actually show. We could not open the text of the amending regulation on EUR-Lex for this article, so we do not describe that mechanism in detail.
  • Standards. The harmonised standards that would give manufacturers a clear way to show compliance are still being drafted. We track this in our article on AI Act harmonised standards at CEN-CENELEC.
  • The Machinery Regulation itself. Separately from the AI Act, the Machinery Regulation applies on a mandatory basis from 20 January 2027, according to the Commission's machinery page. For most robot manufacturers, that is the nearer deadline.
  • Guidance. Commission guidance on how the new safety-component wording applies to specific product types would help, and organisations should watch for it.

What this means for event organisers

  • Nothing new starts tomorrow. The transparency duties and the AI literacy duty already apply. The Omnibus did not add new obligations for companies that hire robots for events.
  • Ask suppliers the right question. Instead of asking whether a robot is "AI Act compliant", ask whether its safety-relevant functions are covered by its machinery conformity assessment and CE marking, and request this in writing.
  • Keep briefing your team. The reworded Article 4 is softer, but it still asks deployers to support the AI literacy of the people operating the system. A short documented briefing for the operators and event staff is a sensible minimum.
  • Watch 2027 and 2028. The Machinery Regulation from January 2027 and the high-risk dates in December 2027 and August 2028 are the milestones to plan contracts and long-term hires around.

Frequently asked questions

When did the Digital Omnibus on AI become law?

The Commission proposed it on 19 November 2025. Parliament and Council reached political agreement on 7 May 2026, Parliament approved the text on 16 June and the Council on 29 June. It was published in the Official Journal on 24 July 2026 as Regulation (EU) 2026/1744 and entered into force on 27 July 2026, according to the Commission and the Parliament's Legislative Observatory.

When do the AI Act high-risk rules now apply to robots?

The Commission lists 2 December 2027 for high-risk systems in the sensitive areas of Annex III and 2 August 2028 for AI built into products covered by EU product law, naming machinery as an example. Because machinery has moved to Section B of Annex I, the AI requirements for robots are meant to come through the Machinery Regulation, and the details of that integration are still to be seen.

Did the Omnibus remove the AI literacy obligation?

No. Article 4 was reworded, not deleted. Providers and deployers must now take measures to support the development of AI literacy among staff and others operating AI systems on their behalf. The Commission and Member States must support these efforts, especially for SMEs. Teams running robots at events should still be briefed on how the system works and where it can fail.

Is a robot's conversational AI a safety component under the amended AI Act?

Usually not. The amended Article 6 says AI used only for non-safety purposes such as user assistance, automation or convenience is not a safety component, unless its failure would endanger health and safety. Balance control, emergency stopping or navigation near people are more likely to count. Whether a specific robot is high-risk depends on its design and should be confirmed by the manufacturer.

How this article was made: written by the humanoid4rent editorial team with AI assistance; every factual statement was checked against the sources listed below on 2026-11-14. Spotted an error? Write to info@humanoid4rent.com.

Zdroje

Fakty v tomto článku boli overené podľa týchto stránok v deň jeho vzniku.

  1. Simplification of the implementation of harmonised rules on artificial intelligence – Digital Omnibus on AI (Omnibus VII), 2025/0359(COD) – European Parliament Legislative Observatory
  2. AI Act | Shaping Europe's digital future – European Commission
  3. AI Omnibus enters into force – European Commission
  4. Annex I: List of Union Harmonisation Legislation – AI Act Service Desk, European Commission
  5. Article 2: Scope – AI Act Service Desk, European Commission
  6. Article 4: AI Literacy – AI Act Service Desk, European Commission
  7. Article 6: Classification Rules for High-Risk AI Systems – AI Act Service Desk, European Commission
  8. Machinery – Internal Market, Industry, Entrepreneurship and SMEs – European Commission

Ďalšie z blogu

Chcete robota na svojom nasledujúcom podujatí?

Popíšte raz termín, miesto a to, čo má robot robiť. Dodávatelia odpovedia ponukami, ktoré môžete porovnať na rovnakom základe.

Odoslať dopyt

Pracujete v robotike? Píšte pre náš blog. →