AI literacy, without the fear

Artificial intelligence has moved quickly from the news pages into ordinary offices. Eurostat reports that 20 per cent of EU enterprises with ten or more employees used at least one AI technology in 2025, up from 13.5 per cent a year earlier. The gap by size is striking: 17 per cent of small firms, about 30 per cent of medium-sized firms and 55 per cent of large ones. In Germany the figure was 26 per cent, in France 18 per cent. In the United Kingdom, the Office for National Statistics estimates that about 35 per cent of firms with ten or more employees were using AI by mid-2026.

What the European rules actually say

The EU AI Act entered into force in August 2024. One of its first provisions, Article 4 on AI literacy, has applied since 2 February 2025. In July 2026 the so-called Digital Omnibus on AI amended the Act. Article 4 now asks providers and deployers of AI systems to take measures to support the development of AI literacy among their staff. The same package moved the main obligations for high-risk systems to December 2027 and August 2028.

For most small and medium-sized companies, the practical message is calm. Using a writing assistant or a scheduling tool is not a high-risk activity. What the rules encourage is sensible: that people who use AI at work understand roughly what it does, where it can go wrong and what they should not put into it.

The United Kingdom has chosen a different, principles-based route without a single AI law. But British firms that serve European customers will meet the same questions in contracts and supplier questionnaires.

A good habit, not a compliance project

AI literacy does not require a training academy. For a firm of fifty people it can be a short internal guide, an hour of practical training and a named person who answers questions. It covers which tools are approved, what information must never be entered, how to check results, and when a human decision is required.

This protects the company, but it also helps people. Staff who understand a tool use it better and trust it for the right things. And owners gain something valuable: a clear view of where AI is saving time, and where it is only creating extra checking.

Start with the work, not the tool

The companies that benefit most start with a task that is slow, repetitive or error-prone, such as drafting quotations, answering routine questions or sorting documents. They test a tool on that one task, measure the result and decide. That approach is cheaper than a large programme, and it builds exactly the kind of understanding the rules have in mind.

Published 1 October 2026 · © 2026 Habsburg Digital Ltd