You run a research team and use four AI tools. The European AI Act sorts every AI use into tiers, from Prohibited at the top to Minimal at the bottom, with a Transparency tier in the middle whose only real duty is to say that AI was used. Place each tool in the tier you think the law puts it, then check your guesses. Most people rank the danger the wrong way round.
Placements are this article's reading of the AI Act text and the European Commission's guidance of 4 February 2025, not an official tool-by-tool map. No standards body has published one, so a real team has to place its own tools.