The European Commission wants to give companies more time to adapt to the EU’s new artificial intelligence rules — but experts and Members of the European Parliament (MEPs) say they regret seeing key obligations delayed so soon after the landmark AI Act was adopted.

On Wednesday, the Commission unveiled its new “Digital Omnibus,” a package designed to streamline and simplify several major EU digital regulations. One of the most significant changes concerns the AI Act, whose rollout is already phased. With the Omnibus, the Commission has now formally proposed delaying some of the strictest requirements.

High-risk AI rules postponed

The Omnibus grants businesses and public bodies using “high-risk” AI systems — tools used for CV screening, grading school exams, or assessing loan applications — more time to comply. Instead of coming fully into force in August 2026, these obligations would now apply starting December 2027, more than a year later than originally planned.

This delay could have substantial real-world consequences. If adopted, the Omnibus would allow AI systems to rely on categories of data that were previously restricted when making decisions about access to essential financial services.

Peter Norwood, Senior Research and Advocacy Officer at Finance Watch, warned that the proposal essentially pursues a “deregulate to accelerate” strategy — with consumers bearing the risks.

“Under these proposals, a person could be denied a loan because of a biased AI model, or charged higher insurance premiums based on predicted health status, all without their knowledge or consent,” he said.

Why push back the timeline?

The Commission attributes the delay to a lack of preparedness from EU member states and the need for companies to adjust to the complexity of the new requirements.

Big tech lobby group CCIA — whose members include Amazon, Apple, Google, and Uber — welcomed the additional time but argued the Commission did not go far enough. They called for “bolder” and “clearer” revisions, including an update to the criteria for identifying AI models that pose “systemic risk” and a fix for copyright-related wording they say conflicts with international standards.

Consumer groups, however, sharply criticized the move.

Agustín Reyna, Director General of the European Consumer Organisation (BEUC), said the proposal appears to prioritise the interests of major tech companies over those of European citizens.

“Consumers were promised simplification to support the European economy, and yet the Commission’s proposal can only be read as deregulation to the almost exclusive benefit of Big Tech,” he said.
“Instead of cutting consumer rights, the EU should focus on making compliance easier for European companies and consumers alike.”

A number of AI experts say the delay is unsurprising, pointing out that many EU countries have still not created the enforcement authorities required under the AI Act.

Hanane Taidi, Director General of TIC Council — representing independent bodies tasked with assessing AI compliance — noted:

“Many member states missed the August 2025 deadline to designate competent authorities. Until these structures are in place, Conformity Assessment Bodies cannot be notified, and the system cannot fully function.”

A politically difficult road ahead

Adopting the Omnibus may prove challenging. The proposed changes would require amending the EU’s General Data Protection Regulation (GDPR), one of the bloc’s most politically sensitive laws. Several MEPs from centrist and left-leaning groups have already expressed firm opposition.

Francine Cunningham, Regulatory and Public Affairs Director at Bird & Bird, told Euronews the Omnibus is likely to face pushback:

“We can expect resistance since the omnibus process bypasses the usual impact assessments and consultations. Some of the laws being amended were only recently adopted and, in some cases, not fully implemented yet. MEPs may be reluctant to change rules they debated and approved just months ago.”

Despite the tensions, the clock is ticking. The legal framework still needs to be completed before August 2026, the original deadline for implementing the rules governing high-risk AI systems.

Source