Synthetic and emotion flags
Two flags that change the notice: synthetic content and emotion recognition.
Two Art. 50 duties are missed more often than the basic "this is AI" disclosure, because they attach to specific modalities rather than to the product as a whole: synthetic content and emotion recognition.
Synthetic content
Content generated or manipulated by AI that resembles real people, places, or events needs disclosure — and, where technically feasible, machine-readable marking. The marking duty sits with the provider; as a deployer your job is to ask and record the answer.
Emotion recognition
Systems that infer emotions need notice to the people exposed — and in workplaces and schools the practice is banned outright, which makes the notice question moot and the deployment question immediate.
Why the flags matter
A chatbot that sounds human fails the basic duty. A synthetic-media feature fails the marking duty. A workplace emotion feature is not a disclosure problem at all. The flags tell you which failure you are looking at.
How the tool frames it
It asks about modality and flags both duties where the facts support them. It drafts the notice themes; design and counsel ship the wording.
Decision-support drafting only — not legal advice. Confirm the duty set against the current consolidated text of Regulation (EU) 2024/1689.