Today marks one of the most significant milestones since the adoption of the EU AI Act.
From 2 August 2026, key provisions of the AI Act become applicable across the EU, marking the transition from legislative preparation to regulatory enforcement.
For organisations developing, deploying or using AI, compliance is no longer simply a future project. It has become an operational and regulatory reality.
A New Chapter for AI Regulation
Today’s milestone represents far more than another compliance deadline.
The European Commission’s AI Office, together with national competent authorities, now begins enforcing key provisions of the AI Act. Following years of legislative negotiations, political agreement, implementing measures and regulatory guidance, the focus now shifts from understanding the Regulation to demonstrating compliance with it.
For organisations across virtually every sector, AI governance is becoming part of day-to-day business rather than a topic reserved for future planning.
Transparency Requirements Become Applicable
Among the most important provisions taking effect today are the AI Act’s transparency obligations.
Depending on the circumstances, organisations must ensure that individuals are informed when they are interacting with an AI system rather than a human. Certain AI-generated or AI-manipulated content must also be appropriately disclosed, while providers of general-purpose AI systems must comply with new requirements concerning the marking of AI-generated content.
These obligations are intended to strengthen transparency and trust by ensuring that people can recognise when artificial intelligence is involved in producing content or interacting with them.
A New Role for the AI Office
Today also marks the beginning of the AI Office’s supervisory role in relation to providers of general-purpose AI (GPAI) models.
As increasingly capable foundation models become embedded across countless digital products and services, European oversight enters a new phase. While many obligations under the AI Act will continue to apply progressively over the coming years, today’s milestone formally launches the AI Office’s operational role under the Regulation.
What Should Organisations Do Now?
The AI Act is entering its implementation phase.
Organisations should ensure that they have identified AI systems within the scope of the Regulation, assessed whether the transparency obligations apply, reviewed governance arrangements and updated internal policies, procedures and user-facing documentation where necessary.
Equally importantly, AI compliance should not be viewed as a one-off exercise. As further provisions become applicable over the coming years, organisations will need to continue monitoring regulatory guidance, enforcement practice and legislative developments.
Looking Ahead
Today’s milestone is only one step in the AI Act’s phased implementation. Additional obligations relating to prohibited AI practices, high-risk AI systems and other aspects of the Regulation will continue to take effect over the coming years.
Nevertheless, 2 August 2026 will undoubtedly be remembered as the date on which the AI Act moved from legislative ambition to practical enforcement.
For businesses, public authorities and AI developers alike, the conversation is no longer about preparing for the AI Act. It is about complying with it.
A personal note
As it happens, I also share my birthday with this important AI Act milestone. I suspect the Regulation will receive rather more attention today – but I’m very happy to celebrate both.
And to everyone working in AI, privacy and digital regulation: Happy AI Act Day! 🎂
