EU Approves Delays and Other Amendments to Certain EU AI Act Obligations: What Businesses Should Know

The European Parliament on 16 June 2026 granted its final approval to certain material amendments to the EU Artificial Intelligence (AI) Act. These amendments include delaying the application of certain significant obligations in relation to high-risk AI systems and simplifying other regulatory rules in the act.

These amendments follow sustained industry pressure and concerns about the European Union’s ability to timely release supporting compliance frameworks and regulatory guidance. The amendments also sit within a wider EU effort to streamline and rationalise digital regulation, including “omnibus”-style measures aimed at reducing overlap and administrative burden.

As a next step, the European Council is expected to formally adopt the amendments, which would thereafter lead to its formal publication in the Official Journal. This process to turn the amendments into law is expected to be complete by 2 August 2026. Until such time, the act, in its current form, continues to be the law.

Overall, we would suggest that businesses treat these developments primarily as an extension of time to complete their AI Act compliance efforts, rather than as a material relaxation of the underlying obligations as such.

HOW THE AMENDMENTS WILL IMPACT EFFECTIVE DATES UNDER THE AI ACT

The amendments will push back the effective date for many obligations applicable to high-risk AI systems. Notably, key obligations applicable to the following:

• Standalone high-risk AI systems within the categories listed in Annex III (e.g., education, employment, critical infrastructure, credit scoring, law enforcement, etc.) will take effect from 2 December 2027 rather than on 2 August 2026 as originally set out in the Act (i.e., 16-month delay)
• High-risk AI systems that are products, or safety components of products, regulated by the EU product safety laws listed in Annex I of the AI Act (e.g., medical devices, radio, toys, etc.) will take effect from 2 August 2028 rather than on 2 August 2027 as originally set out in the Act (i.e., a 12-month delay)

Separately, watermarking obligations for certain AI systems placed on the market before 2 August 2026 will take effect from 2 December 2026 rather than 2 August 2026. Companies will need to continue to consider their transparency obligations under the Act more generally.

WHY THE PROPOSED DELAY?

The primary driver is the slower than expected development of harmonised technical standards and guidance, which are critical to operationalising the AI Act.

European standardisation bodies have faced delays in delivering key standards, with many now expected towards the end of 2026. Without these tools, organisations would face significant uncertainty in determining how to meet their compliance obligations.

As a result, EU policymakers appear increasingly aligned on the need to sequence obligations with the availability of implementation guidance, rather than risk premature enforcement.

WHAT ARE OTHER NOTABLE AMENDMENTS PROPOSED TO THE AI ACT?

Additionally, some of the other main envisaged changes to the AI Act include the following:

• New banned system: Nudifier applications (i.e., those that create non-consensual sexually explicit or intimate content, or child sexual abuse material) are to be categorised as prohibited AI systems from 2 December 2026
• Narrowing of “safety component” definition: In relation to high-risk AI systems, AI systems that are used solely to fulfil functions related to user assistance, performance optimisation, service efficiency, automation, convenience or quality control operations of non-safety related aspects will not be regarded as safety components unless their failure or malfunctioning would endanger health and safety
• New limited special category data use exemption: In exceptional circumstances, where strictly necessary and subject to safeguards, special category data may be processed for bias detection and correction
• Simplified compliance for small mid-cap companies (SMCs): The simplified compliance framework applicable to SMEs will also be extended to SMCs (i.e., those companies with up to 750 employees and €150 million in annual revenue or an annual balance sheet total not exceeding €129 million)
• Simplification of sector-specific legislation: The European Commission will be granted power to issue secondary legislation disapplying specific provisions of the AI Act in relation to high-risk AI systems that are products, or safety components of products, when the legislation listed in Annex I provides an equivalent or higher level of protection and AI systems within scope of the Machinery Regulation ((EU) 2023/1230) will primarily be governed by sectoral rules as opposed to the AI Act (i.e., this regulation is being moved from Section A to Section B of Annex I)

WHAT BUSINESSES SHOULD DO NOW

Despite the anticipated delay and the other proposed amendments, the direction of travel under the EU AI Act remains materially unchanged in most respects.

Organisations deploying or developing AI (particularly in HR, financial services and other Annex III sectors) should continue to prepare by taking the following steps:

• Mapping: Mapping and classifying AI systems in accordance with the AI Act classifications, with a particular focus on assessing potential unacceptable risk and high-risk classifications
• Governance and compliance: Building governance frameworks, including policies, controls and accountability structures and compliance programmes, including developing technical documentation, training programmes and risk management processes, aligned with AI Act requirements
• Monitoring developments: Closely monitoring forthcoming legislative updates, standards, guidance and codes of practice as well as ensuring practices are in place to maintain updated mapping, governance and compliance practices as needed to ensure ongoing compliance

Importantly, organisations should also be aware that the application of certain existing elements of the AI Act, such as prohibitions on specific practices deemed to present unacceptable risks, and (watered-down) AI literacy requirements, remain in effect.

KEY TAKEAWAY

The compliance timeline may be shifting, but the substance of the AI Act is largely not. Businesses that use this period to embed robust AI governance practices will be significantly better placed to manage regulatory risk under the AI Act and demonstrate responsible AI practices as these rules come fully into force.