The European Commission is currently presenting its "Digital Omnibus," a strategy designed to make the EU's extensive digital ...
The European Commission is currently presenting its "Digital Omnibus," a strategy designed to make the EU's extensive digital rules more straightforward.
Included in the legislation is the AI Act, which was approved last year. The law is being put into effect slowly – however, with the Omnibus, the Commission has formally stated that certain provisions regarding the most dangerous forms of the technology will be postponed.
Specifically, the Ominbus provides additional time for businesses and organizations implementing "high-risk" AI technologies, which are applied in areas like reviewing resumes, grading school exams, or evaluating loan applications. These entities will now be exempt from the full scope of the bill's requirements until December 2027, over a year later than the initial date of August 2026.
This might have a significant tangible impact. Should the Digital Omnibus be implemented, AI systems will be capable of utilizing previously restricted data to determine access to critical financial services.
Peter Norwood, a Senior Research and Advocacy Officer at Finance Watch, stated that this represents a "deregulate to accelerate" approach to AI, adding that consumers will eventually bear the consequences.
According to him, under these suggestions, an individual might be refused a loan due to a prejudiced AI system, or face increased insurance costs based on anticipated health conditions, all without their awareness or approval.
Why this delay?
The Commission attributes the delay to the insufficient implementation by member states and the necessity for companies to have more time to adjust to the intricate new regulations.
A major technology industry group, the CCIA, which includes companies like Amazon, Apple, Google, and Uber, supported the proposed postponement in the Omnibus, yet urged for "more courageous" and "more explicit" measures.
"The Omnibus fails to seize important chances to update the outdated compute threshold for recognizing AI models that present a 'systemic risk,' and does not address problematic language regarding the extraterritoriality of copyright rules, which contradicts EU and global principles," the group stated in a release.
Some detractors of the Omnibus, nevertheless, argue that the break extends too much.
"Consumers were promised easier processes to boost the European economy, but the Commission's proposal seems to be seen as a move towards reducing regulations that mainly benefits large technology companies," said AgustÃn Reyna, Director General of the European Consumer Organisation BEUC.
Rather than limiting consumer rights, the European lawmaker ought to concentrate on simplifying adherence, benefiting both European businesses and consumers.
Although France and Germany have already accepted the AI Act's rules, many leading AI experts are disappointed that several EU countries failed to establish the necessary enforcement mechanisms to meet the legislation's original deadlines.
"Several member states failed to meet the August 2025 deadline for appointing designated authorities," stated Hanane Taidi, Director General of the TIC Council, an organization that represents independent firms tasked with evaluating adherence to the AI Act.
Until these national frameworks are established, Conformity Assessment Bodies cannot be officially recognized, and the system will not operate completely.
The next hurdle
The implementation of the Omnibus could face challenges. The suggested amendments to the EU's AI regulations will involve a contentious adjustment to the General Data Protection Regulation (GDPR), and numerous MEPs from both the left and center of the political spectrum have already expressed their disagreement.
Francine Cunningham, who serves as the Regulatory and Public Affairs Director at the law firm Bird & Bird, stated to Euronews: “We should anticipate some pushback in the future because the omnibus process skips the standard impact evaluations and discussions typically needed for legislative proposals.”
Some of the laws under review were recently enacted and, in certain instances, have not yet been fully put into effect. As a result, MEPs might be hesitant to alter legislation they only recently discussed and approved.
Nevertheless, the regulations must still be established prior to the August 2026 deadline—the initial date set for the enforcement of "high-risk" AI rules.
