What does the EU AI Act change for our company?

The European regulation on artificial intelligence, the EU AI Act, adopted in 2024, is coming into force in stages. It does not target AI in general, but its uses, classified according to the risk they pose to people. Some are prohibited. Others, known as high-risk, such as sorting job applications or assessing creditworthiness, are subject to stricter obligations. Most everyday uses have only limited obligations, transparency in particular.

Your company is affected even if it develops nothing: the regulation also sets obligations for those who use these systems. You still need to know which ones you use. That is often where things get stuck: tools bought by a department, AI features quietly added to existing software, individual uses nobody has declared.

• Few companies have an inventory of their uses.Without one, it is impossible to know which obligations apply.
• AI also arrives through existing software.Features appear with updates, with no project and no decision.
• Human oversight must be real.For sensitive uses, you need to be able to show who supervises, how and with what means.

1. 01List your AI uses, including those in software you already use and those of your teams.
2. 02For each one, note what it decides or influences, and which people it affects.
3. 03Have a lawyer check the classification of uses that concern recruitment, credit or access to a service.

We help companies list their uses and document their choices, working with their advisers on legal questions.

• Should we tell our customers they are talking to an AI?→
• Who is responsible when an AI makes a mistake?→
• Is generative AI compatible with the GDPR?→
• How far should AI be allowed to decide on its own?→