In short: The EU AI Act is no longer a topic for the future. Since February 2, 2025, certain AI practices have been prohibited and companies have had to promote AI literacy. Since August 2, 2026, transparency obligations have applied, for example to chatbots and AI-generated content. With the Digital Omnibus, the EU pushed back the obligations for high-risk systems to December 2, 2027, and August 2, 2028, respectively. Advisory and supervisory boards should use this time now to take stock. There is no reason to sound the all-clear.
Why is the AI Act a topic for the board?
The AI Act (Regulation (EU) 2024/1689) doesn’t just affect companies that develop AI. It also places obligations on companies that use AI systems, which the law calls “deployers.” That brings almost every mid-sized company into scope as soon as it uses AI in HR, customer service or production. The fines are substantial.
7%
For prohibited practices, fines can reach €35 million or 7 percent of global annual revenue, whichever is higher.
For a board, this is less about legal fine print than about an organizational question: does management know where in the company which AI is being used, and is it clear who is responsible for it?
Which obligations already apply, and which were postponed?
With the Digital Omnibus on AI (Regulation (EU) 2026/1744), in force since July 27, 2026, the EU adjusted the timeline. Here is where things stand:
| Since / from | What applies | What it means for companies |
|---|---|---|
| February 2, 2025 | Prohibited AI practices (Art. 5) and AI literacy (Art. 4) | Certain applications, such as social scoring or manipulative techniques, are prohibited. Providers and deployers must take measures to build their employees’ AI literacy. |
| August 2, 2025 | Obligations for providers of general-purpose AI models | Mainly affects model developers, not their users. |
| August 2, 2026 | Transparency obligations (Art. 50) | People must be able to tell that they are interacting with an AI system. AI-generated content and deepfakes must be labeled. |
| December 2, 2027 | High-risk systems under Annex III (previously August 2026) | For example, AI in recruiting and workforce management, in creditworthiness assessments of individuals or in critical infrastructure. |
| August 2, 2028 | High-risk systems in regulated products under Annex I (previously August 2027) | For example, AI as a safety component in machinery or medical devices. |
The Omnibus watered down the AI literacy obligation but didn’t abolish it. Companies must take measures to support their employees’ AI literacy. They don’t have to guarantee a specific level of competence for individuals (overview of the Digital Omnibus).
What does the postponement mean for mid-sized companies?
The postponement buys time, but it’s no reason to put the topic off.
Three reasons argue against waiting. First, the prohibitions, transparency and literacy obligations already apply. Second, experience shows it takes a long time for a company to even know which AI systems are in use, especially when business units use their own tools. Third, customers in regulated industries are already asking for proof of traceability and data handling, regardless of legal deadlines. Build that early and you turn an obligation into a sales advantage.
What questions should an advisory board ask about the AI Act?
- Do we have an inventory of all AI systems in use, including the tools business units use on their own?
- Who in the company is responsible for AI Act compliance, and who does that person report to?
- Which of our applications could qualify as high-risk systems, for example in HR or credit checks, and how are we preparing for the 2027 and 2028 deadlines?
- What measures have we taken on AI literacy, and are they documented?
- Do our chatbots and AI-generated content meet the transparency obligations?
- Are we a provider or a deployer? Companies that sell their own AI solutions to customers carry significantly broader obligations.
How do you connect compliance with strategy?
The AI Act is often discussed as a brake. From a board’s perspective, it’s also a framework that forces questions that should be asked anyway: where do we use AI, with what data, under whose responsibility and with what oversight? Companies that can answer these questions usually also use AI more successfully in business terms, because they know what they’re doing. An advisory board should therefore make sure that the same management level owns both compliance and AI strategy.
Frequently asked questions
Does the EU AI Act also apply to companies that only use AI?
Yes. The AI Act distinguishes between providers, who develop AI systems or place them on the market, and deployers, who use them. Deployers have obligations too, for example on AI literacy, on transparency and when using high-risk systems.
Has the EU AI Act been postponed?
In part. The Digital Omnibus on AI, Regulation (EU) 2026/1744, pushed back the obligations for high-risk systems: to December 2, 2027, for systems under Annex III and to August 2, 2028, for systems in regulated products under Annex I. The prohibitions, AI literacy and transparency obligations still apply.
What does the AI literacy obligation require?
Providers and deployers must take measures to support the AI literacy of their employees and of people acting on their behalf. Since the Digital Omnibus, they no longer have to guarantee a specific level of competence for individuals. Penalties for violations are set by the member states.
This article reflects the situation as of October 2026 and does not replace legal advice. For the assessment of specific applications, consult a specialized lawyer.
Further reading: Five questions an advisory board should ask management about AI · Evaluating AI investments · Governance starts with trust
Sources: Regulation (EU) 2024/1689 (AI Act), EUR-Lex · Regulation (EU) 2026/1744 (Digital Omnibus on AI), EUR-Lex · artificialintelligenceact.eu: Digital Omnibus