The European Commission has revealed that a dedicated code of practice to help companies comply with the European Union’s AI Act may not be ready until late 2025, sparking concerns among developers and digital rights groups. The announcement comes as the EU’s landmark AI regulation, passed in March 2024, officially entered into force on July 1, 2025, with phased implementation over the next two years.

Why the Code Matters

The AI Act classifies artificial intelligence systems into categories based on risk, with “high-risk” applications such as biometric identification, recruitment tools, and credit scoring systems facing the strictest oversight. Many companies, especially small and medium enterprises (SMEs), have expressed the need for concrete guidelines to interpret the law’s provisions. A formal code of practice would provide these much-needed clarifications and operational guidance.

Commission’s Timeline and Industry Reaction

Speaking in Brussels, EU officials acknowledged the importance of the code but stated it may not be finalized until Q4 2025. This delayed timeline has raised questions about how businesses—especially those developing or deploying high-risk AI—will meet compliance expectations in the interim. DigitalEurope and other industry groups argue that the absence of guidance could stifle innovation and delay product launches.

Phased Rollout of the AI Act

Although the AI Act is now legally in effect, its obligations will be introduced in stages. The ban on certain prohibited AI systems—such as social scoring and manipulative algorithms—takes effect six months from now. Rules for high-risk systems will apply 24 months from the enforcement date, giving companies time to adapt. However, without a code of practice, many may face uncertainty during the transition.

A Collaborative Drafting Process Ahead

The Commission plans to work closely with member states, industry representatives, AI developers, and civil society to draft the code. It will be part of the AI Pact, a voluntary scheme encouraging early compliance before full enforcement. The Pact has already seen participation from tech leaders such as SAP, Siemens, and IBM, who support its goals but stress the need for timely and practical resources.

What’s Next for Businesses?

Until the code is finalized, the EU recommends that companies begin conducting internal assessments, set up AI governance frameworks, and familiarize themselves with the obligations under the AI Act. Independent regulatory bodies in each EU country will begin offering workshops and compliance clinics later this year to support early implementation.

Conclusion
As Europe sets the global pace on AI regulation, the success of the AI Act will depend not just on rules but on the clarity and support provided to those who must follow them. The code of practice—though delayed—will play a critical role in bridging legal principles with real-world AI development and deployment.