• EU plans to delay high-risk duties in the AI Act until late 2027.
• Officials say firms need more time, but rights groups see weaker protection.

Europe is preparing to slow down parts of its major AI rulebook, giving companies more time to adjust to the strictest parts of the law. The European Commission said this week that it plans to delay the enforcement of high-risk AI rules until December 2027, a move that follows months of pressure from global tech companies and business groups.

The plan is part of a wider effort the Commission calls the Digital Omnibus package. It is meant to simplify several overlapping digital rules without changing the core ideas behind them. During a briefing, a Commission official explained the goal by saying, “Simplification is not deregulation. Simplification means that we are taking a important look at our regulatory landscape,” according to Reuters.

What the delay covers

The changes affect parts of the AI Act, which became law last year and was set to phase in new duties for companies over the next three years. The rules for high-risk systems – like biometric identification, credit scoring, job screening, health services, exams, law enforcement tools, traffic systems, and utility management – were meant to take effect by mid-2026. Under the new plan, full enforcement would not begin until late 2027.

The systems raise concerns because they can influence decisions about a person’s identity, access to services, movement, or legal rights. The Commission argues that companies need more time to meet the technical and legal demands that apply to these tools.

The Digital Omnibus package also includes changes to GDPR, the e-Privacy Directive, and the Data Act. One of the most debated ideas is a GDPR update that would allow companies – including Google, Meta, OpenAI, and others – to use personal data from people in Europe to train AI models. The would rely on clear rules and safeguards, but privacy groups say it may weaken protections that have been in place since 2018.

The push to simplify cookie consent

The proposal also calls for simpler cookie and consent rules. People in Europe face constant pop-ups when they browse the web, and regulators say most users click “accept” without reading anything. The Commission wants to allow quicker choices, reduce repeated consent requests, and set clearer limits on how long cookie settings remain valid.

Supporters argue that this could reduce “banner fatigue” and make websites easier to use. Critics worry that easier consent may open the door to more tracking unless the rules are written carefully.

Why the Commission is doing this

The Commission says the slowdown does not weaken its stance on AI safety. Officials argue that Europe needs a clearer set of rules that companies can actually follow, rather than an ambitious framework that becomes too hard to enforce in practice. The delay ties several obligations to the readiness of technical standards, which are still being developed with industry and regulators.

There is also pressure from outside Europe. Tech firms from the United States and Asia have warned that strict rules could discourage investment or slow down AI research in Europe. Some EU member states share those concerns and want more time for national agencies to prepare.

Business groups have welcomed the extra time, saying the original deadlines were tight and risked confusion. Companies deploying AI in healthcare, banking, hiring, and public services have argued that they need clearer guidance and workable compliance steps.

Strong criticism from rights groups

Digital-rights advocates see the changes differently. Groups like European Digital Rights (EDRi) warn that the package would weaken data-protection rights and make it easier for large companies to gather personal data. They say the delay on high-risk rules may leave people exposed to systems that can deepen bias or make mistakes in sensitive areas like hiring or policing.

Other critics argue that the EU is stepping back from its position as a strict regulator. They point to earlier moments when the Commission eased pressure on environmental rules after industry pushback and say this digital shift follows the same pattern.

What happens next

The Digital Omnibus package still needs approval from the European Parliament and the Council. Lawmakers are expected to debate the proposal in the coming months, and some may push for changes before voting. Even after it passes, the exact start dates for some rules depend on when supporting standards are ready, which adds another layer of uncertainty for companies.

For now, Europe is trying to balance two goals: protect people from risky AI systems and keep its tech sector competitive. The months ahead will show whether the new plan can do both.

Want to experience the full spectrum of enterprise technology innovation? Join TechEx in Amsterdam, California, and London. Covering AI, Big Data, Cyber Security, IoT, Digital Transformation, Intelligent Automation, Edge Computing, and Data Centres, TechEx brings together global leaders to share real-world use cases and in-depth insights. Click here for more information.

TechHQ is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.