The EU AI Act was once championed as the gold standard for global AI governance but is now facing calls for delay from some of its own leaders, most recently and notably, Swedish Prime Minister Ulf Kristersson.

Kristersson called the Act “confusing” and suggested that its rollout be paused until common technical standards are in place. “An example of confusing EU regulations is the fact that the so-called AI Act is to come into force without there being common standards,” he told the Swedish parliament.

While officials from countries like Poland and the Czech Republic have expressed openness to delay, Kristersson is the first EU head of government to formally voice support for a pause, a move applauded by some and criticized by others.

The concern

At the heart of the debate is this: How can companies comply with a regulation when the detailed technical standards to guide that compliance haven’t been finalized?

The AI Act, especially for high-risk and general-purpose AI systems, relies on harmonized standards to give companies a presumption of conformity. These standards, covering cybersecurity, data quality, human oversight and more, are meant to act as a compliance roadmap. Without them, some argue, businesses are left navigating blind.

Henna Virkkunen, the European Commission’s tech chief, has acknowledged this gap. “If we see that the standards and guidelines are not ready in time, we should not rule out postponing some parts of the AI Act,” she said during a meeting of EU digital ministers in Luxembourg.

The Code of Practice for General Purpose AI, which is one of the cornerstones of the AI Act’s rollout, was originally due in May but is already running more than a month behind schedule.

But while the political temperature rises, it’s important to note that the absence of finalised technical standards does not mean companies can’t comply with the AI Act. Lawmakers built in mechanisms precisely for this situation. One of the most important is the introduction of the Codes of Practice, which is supposed to bridge the gap until formal standards arrive. These codes can guide providers of general-purpose AI models in meeting their obligations under the Act.

The legislation is also designed to be functional without standards. Standards help smooth compliance, but they are not a prerequisite. Companies can still meet legal obligations by directly demonstrating conformity with the essential requirements of the Act.

The Washington effect?

Kristersson’s remarks along with the broader pushback comes as Washington’s influence on global tech regulation is growing again. Where once the “Brussels effect” led countries to adopt or align with the EU’s strong digital governance model, we may now be seeing the emergence of the “Washington effect,” – a call for deregulation, delay and simplification.

This shift in narrative is already spreading. Industry leaders are lobbying hard for a “stop-the-clock” mechanism that would delay compliance dates until guidance and standards are fully in place. And EU ministers like Poland’s Dariusz Standerski have echoed this sentiment. “Delaying deadlines alone is not the way,” he said. “First, we need a plan.”

What does this mean for your businesses?

For now, the AI Act is still law and still in effect. Its phased implementation continues, with key compliance dates looming in August 2025 and 2026. Some provisions, such as bans on certain types of AI applications, are already in force.

Businesses developing high-risk or general-purpose AI systems should not count on a pause. Instead, they should:

• Monitor Codes of Practice as they emerge 
• Document their compliance approach even in the absence of finalized standards 
• Engage with regulators and standardization bodies early to stay ahead of evolving expectations 

The EU AI Act was always going to be a bold experiment in global tech governance. Now, it faces its first real test: Can it remain effective and credible in the face of political skepticism, incomplete standards and rising global competition? The coming months will reveal whether Europe doubles down on its vision for trustworthy AI or hits the brakes to recalibrate. Either way, businesses should prepare not just for compliance but for a regulatory landscape that is as dynamic as the technology it seeks to govern.

Want a guide to everything you need to know about the EU AI Act? Get it here.