August 2, 2026: The AI rule your business can't postpone
AI Omnibus may delay some AI Act requirements but transparency rules are still coming in 2026
The European Commission’s Digital Omnibus on AI is designed to simplify the implementation of the AI Act and reduce unnecessary compliance burdens for businesses. It introduces measures to make compliance more practical and proportionate in relation to high-risk AI systems.
It is therefore easy to assume that businesses can afford to postpone their AI compliance projects. But some of the AI Act’s most relevant obligations for everyday business use remain unchanged and are approaching quickly. One of those areas is transparency.
While the Digital Omnibus on AI focuses largely on simplifying parts of the high-risk AI regime, the AI Act’s transparency requirements remain on track and will start applying from 2 August 2026.
For many organisations, these obligations are likely to become relevant long before they need to determine whether they are operating a high-risk AI system.
Transparency rules are coming in August 2026
The transparency obligations under Article 50 AI Act will apply from 2 August 2026.
To help organisations prepare, the European Commission published practical guidance on the labelling of AI-generated content, providing a clearer picture of what compliance may look like in practice. This guidance is expected to become an important reference point for businesses preparing for the new requirements.
These rules are not aimed solely at AI developers. Any organisation using generative AI in customer communications, content creation or public-facing activities should consider whether the transparency requirements may apply.
But the AI Act does not create a general obligation to label every piece of AI-generated or AI- altered content.
The principle is simple: where AI creates or alters content in a way that could be mistaken for something real, people should be informed that AI was involved.
The AI Act therefore focuses on specific situations where individuals could be misled about the nature of the content they are receiving or interacting with.
AI Systems interacting with individuals
If customers or users are interacting with an AI system, they should generally be made aware of that fact, unless it is already obvious from the context.
This requirement is particularly relevant for customer-facing chatbots, virtual assistants and similar AI-powered interfaces.
For example: “You are interacting with an AI-powered virtual assistant.”
Synthetic audio, image and video content ("deepfakes")
The AI Act also contains transparency requirements for AI-generated or AI-manipulated audio, image and video content that could reasonably be mistaken for authentic or real-world content.
This includes AI-generated or AI- manipulated photographs or videos of realistic individuals, AI- generated or AI -manipulated voice recordings and other synthetic media depicting persons, objects, places or events in a way that appears authentic.
Businesses using this type of content should consider whether viewers need to be informed that the material was generated or manipulated using AI.
Examples may include:
– “This image was generated using artificial intelligence.”
– “This video contains AI-generated or AI-modified content.”
Exceptions does exist in certain low-risk uses for minor AI-assisted modifications and content that is clearly artistic, fictional or satirical in nature, where there is limited risk of audiences being misled.
AI-Generated text and public interest content
The rules are less straightforward when it comes to AI-generated text. The AI Act does not require businesses to label every document, marketing text, presentation or
report drafted with the assistance of AI.
Instead, the transparency requirements focus on specific situations involving AI-generated text published for the purpose of informing the public on matters of public interest.
Businesses involved in areas such as news, public information, public health, public safety or political communications should therefore keep a close eye on guidance from the European Commission.
Why businesses should prepare now
There are still open questions around how some of these requirements will work in practice, but that should not stop organisations from preparing now.
Everyone should be identifying use cases involving customer-facing AI interactions, synthetic audio, image and video content and AI-generated content intended to inform the public on matters of public interest.
The challenge will not be the legal obligation itself but creating internal processes that allow teams to identify when transparency obligations are triggered and ensure that disclosures are applied consistently.
High-risk AI classification: another area to watch
Alongside the Digital Omnibus on AI, the European Commission has also published draft guidelines on the classification of high-risk AI systems that provides practical examples of when AI systems should and should not be classified as high-risk. It confirms that not every AI system used in HR, financial services, education or other regulated sectors will automatically fall within the high-risk regime.
Creating an AI inventory and properly classifying AI systems will therefore be one of the most important steps in preparing for compliance.
What businesses should be doing now
Rather than waiting for further legislative developments, organisations should consider identifying AI tools currently used across the business, creating and maintaining an AI inventory, assessing which systems may fall within the scope of AI Act obligations, implementing AI governance and approval processes, reviewing how AI-generated content is created and published, preparing labelling and disclosure procedures where required and training employees on the responsible use of AI. Final Thoughts
While the AI Omnibus may simplify certain aspects of AI Act compliance, transparency obligations remain firmly on track.
For many organisations, reviewing how AI-generated content is created, used and disclosed will be a more immediate priority than determining whether a system falls within the high-risk AI regime.
With further guidance on both transparency and high-risk classification expected over the coming months, now is the right time to build an AI inventory, review AI use cases and put governance processes in place.