Current State of the EU AI Act

Requested Amendments to the EU AI Act

The EU AI Act was adopted on May 21, 2024, when the Council of the European Union formally approved the text previously negotiated with the European Parliament. Upon publication, the EU AI Act was promulgated in the Official Journal of the European Union (L Series) on July 12, 2024, as Regulation (EU) 2024/1689, and entered into force on August 1, 2024.

Following publication and entry into force, a number of challenges emerged over time. In particular, the lack of harmonized standards and delays in the issuance of guidance by the European Commission jeopardize the timely implementation of the obligations set out in the AI Act.

As early as November 19, 2025, the European Commission presented its proposal for a “Digital Omnibus on AI” (COM(2025) 836), aimed at simplifying requirements and postponing certain deadlines under the AI Act. A concise interpretation of the Commission’s position can be summarized as follows: “Making the AI Act more workable through technical optimization and consolidation.”

Since late 2024, the European Parliament has increasingly voiced doubts and has called more forcefully for amendments to the AI Act, arguing that:

• practical implementation reveals weaknesses,
• SMEs and start‑ups could be overburdened,
• AI technology is evolving faster than the law,
• overlaps exist with other digital legislation, and
• Europe’s competitiveness must not be put at risk.

Since March 2026, the European Parliament has been formally calling for amendments to the AI Act, although it does not substantively follow the AI Act Omnibus approach. In summary, the European Parliament’s position may be interpreted as follows: “The AI Act is well‑intentioned, but in some areas unfair, unclear, or insufficiently robust—we want to refine and rebalance it.”

The Council of the European Union (Council of Ministers) also supports amendments to the AI Act, while rejecting any fundamental renegotiation or weakening of the existing level of protection. The Council’s position was formally adopted on March 13, 2026, and serves as its negotiating mandate for the trilogue process. The statement “The AI Act above all must work; without realistic deadlines and clear rules, it cannot be enforced,” can be regarded as a concise summary of the Council’s position.

A comparative overview of the requested amendments from the perspectives of the European Commission, the European Parliament, and the Council can be found in the following table.  

The Digital Omnibus

As proposed by the European Commission in November 2025, the Digital Omnibus was intended to optimize the AI Act and other existing regulatory frameworks, such as the GDPR.

With regard to the EU AI Act, the Digital Omnibus was designed to make AI regulation more practical and innovation‑friendly. It includes adjustments aligning the application deadlines for high‑risk AI obligations with the availability of relevant standards and support tools, expands certain simplifications for SMEs, and seeks to streamline governance by strengthening oversight through the EU AI Office. The Commission’s overarching objective was to simplify and consolidate the EU’s digital regulatory framework. All proposed amendments can be viewed here: Digital Omnibus on AI Regulation Proposal.

Examples of proposed Digital Omnibus reforms in the areas of data, cybersecurity, and privacy include a shift from the current cookie opt‑in model to an opt‑out system, an extension of the data breach notification deadline from 72 to 96 hours, and the creation of a central reporting point for cybersecurity incidents. All proposed Digital Omnibus amendments can be found here: Digital Omnibus Regulation Proposal.

The Trilogue

The purpose of the trilogue negotiations is to reach a political compromise. Trilogue negotiations take place between the European Commission, which submits the original legislative proposal and adopts a mediating role; the Council of the European Union, representing the governments of the Member States; and the European Parliament, representing directly elected Members.

On the evening of April 28, 2026, the trilogue negotiations (provisionally) failed. Despite eleven hours of deliberations, no agreement was reached between the parties on the “Digital Omnibus” reform package. The next round of negotiations is expected to take place in May 2026.

What does this mean in practical terms for companies placing AI systems on the European market? Should providers and deployers simply pause and wait for more concrete guidance?

Preparing with Proportion and Care

For providers, deployers, and other individuals or organizations involved in placing AI systems on the European market, now is the time to take action or to continue doing so.

While the original rules of the AI Act for high‑risk systems are scheduled to enter into force on August 2, 2026, some parliamentary factions are calling for a postponement until the end of 2027 or even 2028. However, these demands have not (yet) been enshrined in law. Important: These calls for postponement are political in nature; the legally binding deadline remains August 2, 2026. This creates a clear challenge for companies and organizations that must prepare for an approaching deadline amid significant temporal and substantive uncertainty.

It is therefore essential to use the current period to prepare for the upcoming requirements. From a forward‑looking perspective, the stalled trilogue process is likely to result in delays and targeted adjustments—but not in a fundamental departure from the EU AI Act. Important: Despite political blockades, the core structure of the AI Act remains intact; companies must continue to align with the applicable version of Regulation (EU) 2024/1689. Compliance efforts should therefore continue to be guided by the existing text of the AI Act. Concrete steps include:

• taking an inventory of all AI systems,
• classifying identified systems by risk,
• performing a gap analysis,
• establishing governance structures, and

ensuring comprehensive documentation.

How to Proceed

Even if the challenge appears significant, developing an organization‑specific AI Act roadmap provides valuable orientation. By identifying regulatory requirements as well as the financial and time resources required, organizations can achieve planning certainty even in a difficult environment. We support this process with a specific advisory offering, the AI Act Roadmap.

For AI providers and deployers, achieving compliance with the AI Act is essential. We support providers in implementing tailored processes and documentation, as well as in managing accompanying communication with notified bodies and competent authorities with our AI Act Compliance for Providers consulting service.

We support deployers, among other things, in implementing their specific obligations through suitable internal processes and in building AI‑related competencies for operational teams: AI Act Compliant Deployment.