Article 4 has required AI literacy since February 2025. Since 2 August 2026, regulators can ask any company that uses AI to prove it, and that deadline didn’t move with the Digital Omnibus.
On 10 September, EU AI Act lawyer Silvia Calls (Abogados Fintech) joined Laura Fernández, CEO of allWomen, and Maida Gómez, Strategic Partnerships at allWomen, for a live panel on what compliance actually looks like. Four things stood out:
- You’re almost certainly a deployer. If your team uses AI tools at work, Article 4 applies, whether or not AI is part of your product.
- High risk depends on impact, not sector. What counts is how a system affects people’s rights and personal data, as in credit or insurance decisions.
- GDPR-style training won’t do. One generic session for everyone falls short. Training has to match the risk each role handles, with evidence that it happened.
- Shadow AI is the gap most companies miss. You can’t train people on, or inventory, tools you don’t know they’re using.
In the recording, Silvia walks through the five things an inspector will ask to see, how inspections usually get triggered and what non-compliance can cost. Maida explains how to build a literacy programme that leaves audit-ready evidence, in courses of 4 to 6 hours.
Watch the full panel
Leave your details and we’ll take you straight to the recording.