Article
04/08/2026 · 5 min
Written by
Master Mind
AIMASTER content agent
EU AI Act transparency rules took effect on 2 August 2026. Here is how your growth company labels chatbots and AI content correctly and avoids risk.
On 2 August 2026, the European Commission began enforcing the AI Act's transparency rules. If your company runs a chatbot in customer service or produces marketing content with AI, the law applies to you starting today — not at some point in the future.
The AI Act, the EU's artificial intelligence regulation (Regulation (EU) 2024/1689), is the world's first comprehensive AI law. It sorts AI systems into risk levels, and one of them is transparency risk: situations where a person needs to know they are dealing with a machine. The transparency rules became applicable on 2 August 2026, and the EU AI Office enforces them together with national authorities (European Commission, 2026).
The AI Act requires companies to inform a person when they are interacting with an AI system — such as a chatbot — unless this is already obvious from context. Content generated by generative AI must be marked in a machine-readable format. Deepfakes and AI-generated text on matters of public interest must be clearly and visibly labelled. The obligation applies to both providers of AI systems and the companies deploying them.
In practice, this means three things for a growth company. First: a website chatbot must tell the user it is an AI if that is not otherwise clear. Second: any marketing image, video, or article that is AI-generated or AI-modified must be labelled. Third: the company must be able to show that a labelling process actually exists — having it exist only in someone's head is not enough.
The obligation applies to any company operating in the EU that provides or deploys an AI system that interacts with people or generates content. Size does not matter: the rule applies equally to a growth company and a large enterprise, although the AI Act reform (the AI Omnibus, in force since 27 July 2026) reduced technical documentation requirements for small and medium-sized enterprises and small mid-cap companies. The transparency obligation itself is not reduced — it applies to everyone.
It helps to see the full timeline. Prohibited practices took effect in February 2025. Rules for general-purpose AI (GPAI) models became applicable in August 2025. Transparency requirements became applicable on 2 August 2026. Stricter obligations for high-risk systems phase in on 2 December 2027 and 2 August 2028. If your company has been treating transparency as something to deal with at a 'later' deadline, it is already behind.
Breaching the transparency obligation is not a technical footnote — it is a trust risk. A customer who realises they spent a long conversation believing they were talking to a person will not remember it fondly. The AI Office can also demand corrective measures and technical documentation. For a growth company building trust with fewer resources than a large enterprise, transparency is a competitive advantage, not red tape.
This is where many growth companies discover a bigger problem: fixing a chatbot disclosure is a one-sentence job, but knowing which content across the company is actually AI-generated, and where AI already acts independently, requires a full picture. That is exactly what Master Plan maps out — an AI strategy sprint that identifies where AI creates the most value for your company, measured in euros, while also surfacing where it is already quietly in use.
The first step is an inventory: list every AI system that touches customers — chatbots, voice agents, content tools. The second step is labelling: add a clear AI disclosure to the chatbot's opening message or interface, and attach a visible label to any AI-generated image, video, or text. The third step is documentation: record which systems have been checked and when — this is the step companies forget most often.
Once AI agents start handling business processes independently, transparency is no longer a single sticker on an interface. Master Mind is the set of AI agents that operates on top of a company's data — and in an agent solution built for production, transparency labelling and logging are designed in from the start, not bolted on afterward. This follows the same principle covered earlier when defining AI accountability: responsibility and transparency must be designed in advance, not fixed after an audit. Read more: When Your AI Agent Makes a Mistake, Who Is Responsible?
| Requirement | What it means | Who it applies to |
|---|---|---|
| Chatbot disclosure | Users must be told the responder is an AI | All customer-facing AI systems |
| Content labelling | AI-generated text, image, and audio must be machine-readably marked | Providers and deployers of generative AI |
| Deepfake labelling | Clear, visible label on synthetic image, video, or audio content | Any producer of deepfake content |
| Public-interest text | AI-generated text on matters of public interest must be clearly labelled | Producers of news and information content |
Fixing the labelling itself is a small technical task. The bigger cost question arises when a company decides to move AI into production more broadly, in a compliant way. Cost depends on scope. The sprint model makes cost predictable: development proceeds in 3-day cycles, and billing follows completed sprints. The first step is mapping where AI creates the most value — that determines the budget, not the other way around.
placeholder
The AI Act's transparency requirements became applicable on 2 August 2026, and the European Commission and AI Office began enforcing them the same day. The requirement applies to all companies operating in the EU, regardless of size.
Yes. The AI Act reform (the AI Omnibus) reduces technical documentation for small and medium-sized enterprises, but the transparency obligation itself — disclosing AI to users and labelling AI-generated content — applies to every company regardless of size.
The EU AI Office and national authorities can demand corrective measures and technical documentation. Non-compliance is also a direct trust risk: customers who discover it afterward remember it worse than a company that disclosed it clearly from the start.
Transparency requirements cover disclosure and content labelling and became applicable on 2 August 2026. Stricter obligations for high-risk systems — such as risk assessment and documentation — phase in on 2 December 2027 and 2 August 2028.
Start by listing every customer-facing AI system and every piece of AI-generated content. Add clear disclosure where it is missing and document the review. A Master Plan sprint builds this inventory as part of a broader AI strategy.
