Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchThe European Union is no longer merely considering a delay. Regulation (EU) 2026/1744, the Digital Omnibus on AI, was adopted on July 8, 2026, published on July 24 and entered into force on July 27. It moves certain high-risk AI obligations, but it does not suspend the AI Act or create a general compliance holiday.
The short answer
The amendment changes the application dates for two main categories of high-risk AI systems:
| System category | Before the amendment | Current application date |
|---|---|---|
| Stand-alone high-risk systems under Article 6(2) and Annex III | August 2, 2026 | December 2, 2027 |
| High-risk systems embedded in products covered by EU product-safety legislation under Article 6(1) and Annex I | August 2, 2027, or the applicable earlier timetable | August 2, 2028 |
| Other AI Act provisions | Varied | Generally unchanged unless the amendment specifically says otherwise |
The dates and categories are set out in the Council’s final legislative summary and the adopted regulation (Council press release; Regulation (EU) 2026/1744).
What was actually delayed?
Stand-alone high-risk systems
The December 2, 2027 date covers many systems listed in Annex III, including tools used for recruitment and worker management, education assessment and admissions, biometric identification or categorisation, critical infrastructure, access to essential private or public services, law enforcement, migration, asylum and border control.
Free tools Windows power users keep installed
One-click scans. No signup required.
#1 Best Overall
High-risk systems in regulated products
The August 2, 2028 date applies where the AI system is a safety component of, or is itself a product covered by, EU harmonisation legislation. Examples can include certain medical devices, toys, lifts, machinery and watercraft. These systems interact with existing product-safety conformity procedures, which is why the timetable is separate.
“High risk” is a legal classification, not a general description of a powerful or dangerous model. The result depends on the system’s intended purpose, the category in Article 6 and the relevant annex, and whether it is connected to a regulated product. A general-purpose or generative model is not automatically a high-risk system.
What the Digital Omnibus changed besides the dates
- Content prohibitions: the package prohibits AI systems used to generate non-consensual sexual or intimate content and child sexual-abuse material.
- High-risk database registration: providers that regard a system as exempt from the high-risk category may have to register that assessment in the EU database.
- Bias testing: the amended framework changes how special-category personal data may be processed for detecting and correcting bias, subject to the regulation’s safeguards.
- Transparency implementation: the timetable for some solutions concerning artificially generated or manipulated content is shortened or adjusted. This does not mean that every transparency duty moved to the same date.
- Supervision and sandboxes: responsibilities and procedures for regulatory sandboxes and national supervision are revised.
- Product-law overlap: the text is intended to reduce duplication between AI Act requirements and sector-specific product-safety legislation.
The May 7, 2026 announcement was a provisional agreement. The final legal position is the adopted and published regulation, not the earlier Commission proposal or an intermediate parliamentary summary.
Why did the EU change the timetable?
The formal rationale focuses on implementation capacity. The EU institutions cited delays in technical standards, national governance arrangements, conformity-assessment capacity and practical guidance. Moving the dates is intended to let those support structures develop while simplifying compliance without reducing the stated protections for health, safety and fundamental rights. The adopted text is available on EUR-Lex.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Rank #3
The change also followed political and commercial pressure. Earlier reporting connected the proposal with lobbying by US technology companies, the Trump administration and European industrial groups concerned about regulatory cost and competitiveness (TechRepublic). That context explains the debate, but it is not proof that foreign pressure alone caused the final law.
What still applies?
The AI Act entered into force on August 1, 2024 and applies in stages. The high-risk postponement does not remove obligations that already apply or that have separate dates. Companies should account for:
- Prohibited practices: bans on practices specified by the Act remain in force according to their applicable dates.
- AI literacy: providers and deployers must take measures to ensure an appropriate level of AI literacy for staff and other people operating systems on their behalf.
- General-purpose AI: important obligations for providers of relevant GPAI models began applying on August 2, 2025; the high-risk delay is not a delay of GPAI duties.
- Transparency: duties for specified AI-generated or manipulated content continue, although particular implementation arrangements have been amended.
- Governance and records: documentation, registration, logging, cooperation with authorities and supervision provisions apply where their own trigger dates have arrived.
- Existing systems: transitional rules can bring an existing high-risk system into scope after a significant design change. The AI Act Service Desk explains these rules in Article 111 guidance.
The European Commission’s staged implementation guidance is available in Navigating the AI Act.
Which businesses and products are most affected?
The revised dates matter most to organizations that build, supply or deploy systems in legally listed high-impact contexts:
- Recruitment, promotion, dismissal and worker-management tools.
- Educational assessment, admissions and examination systems.
- Biometric identification and categorisation.
- Management of critical infrastructure.
- Credit, insurance or other access to essential private or public services.
- Law-enforcement applications.
- Migration, asylum and border-control systems.
- AI safety components in regulated products such as certain medical devices, toys, lifts, machinery and watercraft.
A company can occupy several legal roles at once—for example, product manufacturer, provider and deployer. A supplier’s claim that a product is “AI Act compliant” does not automatically transfer the deployer’s responsibilities.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What companies should do during the transition
- Build an inventory. List every AI-enabled product, internal tool, model and third-party service, including experimental and embedded systems.
- Assign legal roles. Identify whether the organization is acting as provider, deployer, importer, distributor, product manufacturer or more than one.
- Describe intended purpose. Record the real business use, users, affected people and decisions supported by the system.
- Classify the system. Test the intended purpose against Article 6 and the relevant annex, and document why a system is or is not high risk.
- Separate product pathways. Determine whether the system is embedded in a product governed by EU product-safety law, which may place it on the August 2, 2028 track.
- Collect evidence. Review data governance, testing, bias controls, logging, human oversight, cybersecurity, technical documentation and post-market monitoring.
- Review suppliers and contracts. Specify who supplies model information, incident data, evaluations and updates, and do not rely solely on marketing statements.
- Track implementation. Monitor harmonised standards, Commission guidance, national authorities and conformity-assessment capacity.
- Reassess changes. A significant design change to an existing system can affect its transitional treatment.
What the delay does not mean
- It is not a repeal of the AI Act.
- It is not a blanket exemption until 2027 or 2028.
- It does not move every AI Act obligation to one common date.
- It does not make every generative-AI product high risk—or remove obligations that apply to GPAI, prohibited practices, literacy or transparency.
- It does not allow the Commission to postpone the amended deadlines indefinitely; the regulation establishes fixed dates and conditions.
- It does not displace GDPR, product-safety law, employment law, consumer law or national rules that may apply independently.
How the story reached this point
| Date | Event |
|---|---|
| August 1, 2024 | AI Act entered into force. |
| August 2, 2025 | Important GPAI provisions began applying. |
| November 19, 2025 | Commission proposed the Digital Omnibus on AI. |
| May 7, 2026 | Parliament and Council reached a provisional agreement. |
| June 29, 2026 | Council gave final approval. |
| July 8, 2026 | Regulation (EU) 2026/1744 was adopted. |
| July 24, 2026 | The regulation was published in the Official Journal. |
| July 27, 2026 | The amending regulation entered into force. |
| December 2, 2026 | Current official materials identify this as a revised date for certain artificial-content transparency solutions; the exact scope depends on the final text and transitional rules. |
The Council maintains a current overview at its AI Act timeline. Official summaries have described some implementation details differently at different legislative stages, so companies should rely on the final regulation and consolidated guidance when a date affects a filing or launch.
The Bottom Line
The EU’s AI Act delay is now law, but it is a targeted change to the rollout of high-risk rules: December 2, 2027 for many stand-alone Annex III systems and August 2, 2028 for high-risk AI embedded in regulated products. Organizations should use the extra time to classify systems, assign roles and build evidence—not to defer compliance until the deadline.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Recommended Free Tools
