# AI Literacy Has Been Mandatory Since February 2025: How Growth Companies Meet AI Act Article 4 > EU AI Act Article 4 has required sufficient AI literacy from staff since February 2025. Here's how growth companies meet the requirement in practice. - Published: 2026-08-08 - Author: Master Mind - Canonical: https://aimaster.fi/en/artikkelit/ai-lukutaito-on-pakollista-helmikuusta-2025-nain-kasvuyritys-tayttaa-ai-actin-ar 2 February 2025. That's the day Article 4 of the EU AI Act entered into force — long before the August 2026 transparency rules many companies are only now preparing for. Article 4 requires every organization that uses or provides AI systems to ensure staff and other users have sufficient AI literacy. The requirement has been in force for a year, yet few growth companies have documented how they meet it. Under the regulation, AI literacy means the skills, knowledge and understanding that allow people to deploy AI systems in an informed way, aware of their opportunities, risks and potential harms. It isn't a certificate. It's the ability to show that your staff knows when to question an AI system's output. ## Why does this apply to your growth company? The AI literacy requirement applies to every company using AI systems — not just providers of high-risk systems. If your company runs a customer service chatbot, an AI agent handling invoices, or a language model for content, Article 4 applies to you. 57% of Finnish SMEs use AI (Statistics Finland, 2025) — and every one of them should be able to show that users understand the limits of their tools. The problem usually isn't a total lack of knowledge. It's that knowledge sits scattered inside individual employees' heads, never written down anywhere. When a regulator or a customer asks how AI literacy is ensured, "we use it a lot" is not an answer. ## What does AI literacy mean in practice? In practice, AI literacy means three things: staff understand the basic operation of the AI system they use, they can recognize its typical failure modes, and they know who owns the final decision. The required depth differs by role — a developer and a customer service agent need different levels — but both need enough understanding for their own context. The regulation doesn't specify exact training hours or course content. It leaves it to the company to judge what is "sufficient," based on staff's technical background, education, experience, and who the AI system is used on. That flexibility is also a risk: without an active interpretation, the requirement is easy to leave unmet entirely. ## How does a growth company build documented AI literacy? The first step isn't buying a training course — it's mapping: which processes use AI, who uses it, and what risks are involved. That's the same starting point we use in the [Master Plan](https://aimaster.fi/tuotteet/master-plan) sprint — an AI strategy sprint that maps where AI creates the most value for your company, measured in euros. The same mapping also reveals where your AI literacy risk is greatest. After mapping, a company needs three documented elements: role-specific competence requirements, evidence of completed onboarding, and a process for bringing new AI tools under the AI literacy umbrella before rollout. Once a company's data and systems sit on top of [Master Layer](https://aimaster.fi/tuotteet/master-layer) — the data foundation layer that connects CRM, ERP and documents securely for AI use — it also becomes easier to target access rights and training requirements to the right people. The same logic repeats when a growth company deploys AI agents. In our article on [ISO/IEC 42001 certification](https://aimaster.fi/artikkelit/isoiec-42001-sertifikaatti-kannattaako-kasvuyrityksen-hankkia-se-ennen-kuin-asia), we covered AI governance systems more broadly — AI literacy is one concrete component of that, not a separate project. ## What happens if the requirement isn't met? The AI literacy requirement itself isn't covered by the same dedicated penalty provision as prohibited practices, but insufficient AI literacy raises other risks: flawed decisions nobody questions, and a weaker position when a customer or regulator asks about your governance model. In practice, AI literacy is often the first question raised when anyone assesses a company's AI risk. ## Should you build AI literacy in-house or bring in outside help? Both work, as long as the result is documented and demonstrable. Building in-house suits companies that already have a named AI owner and clear processes. An outside mapping is faster when AI use is scattered across many teams and no one has the full picture. Either way, the test is the same: an answer in euros and roles, not slides. ## Frequently asked questions ## When did Article 4 of the AI Act enter into force? Article 4 entered into force on 2 February 2025 (source: artificialintelligenceact.eu). It is one of the first applicable requirements of the AI Act and applies to every organization providing or deploying AI systems in the EU — not only high-risk systems. ## Does the AI literacy requirement apply to small growth companies? Yes. The requirement isn't tied to company size but to whether AI systems are used or provided. If a company uses a chatbot, an AI agent, or a language model in any business process, Article 4 applies the same way it applies to a larger company — only the required level scales with the context of use. ## Is one-off training enough to meet the AI literacy requirement? Not on its own. The regulation requires a "sufficient" level relative to the context of use, and that context changes as a company adopts new AI tools or processes. A working model is a continuous, role-specific practice that updates every time a new system or agent goes into production. ## How does a company know if its AI literacy is sufficient? A clear sign of sufficiency is that staff can explain what task the AI is used for, its typical failure modes, and who makes the final decision. If the answers to these questions are missing or vary by person, documentation and training need an update. AI literacy isn't a separate compliance project you complete once and forget. It's part of the foundation a working AI strategy is built on. Start by mapping where AI is used and who uses it — the rest follows. [Book a free Master Mind analysis](https://aimaster.fi/analyysi) to find out where your company's AI literacy risk is greatest and how to fix it in practice. ## Frequently asked questions ### When did Article 4 of the AI Act enter into force? Article 4 entered into force on 2 February 2025 (source: artificialintelligenceact.eu). It is one of the first applicable requirements of the AI Act and applies to every organization providing or deploying AI systems in the EU. ### Does the AI literacy requirement apply to small growth companies? Yes. The requirement isn't tied to company size but to whether AI systems are used or provided. The required level scales with the context of use, but the obligation applies to everyone. ### Is one-off training enough to meet the AI literacy requirement? Not on its own. The regulation requires a sufficient level relative to the context of use, and that changes as a company adopts new AI tools. A working model is a continuous, role-specific practice. ### How does a company know if its AI literacy is sufficient? A clear sign is that staff can explain what task the AI is used for, its typical failure modes, and who makes the final decision.