Skip to main content
Praxikon
Biometrics

Annex III point 1 - Point 1(c)

Emotion recognition

AI systems for emotion recognition are high-risk (Annex III point 1(c)). In the workplace and in education, emotion recognition is prohibited, except for medical or safety reasons (Article 5(1)(f)).

What is and is not allowed at work and in education is explained at emotion recognition in the workplace: the AI Act ban.

Examples to assess

Emotion analysis of customers in service calls or at a service desk.

Inferring the mood or frustration of visitors from face or voice data, outside work and education.

Classification path

First confirm whether the system is an AI system under Article 3(1).
Determine whether the intended purpose falls within Point 1(c).
Assess whether the Article 6(3) filter may apply and whether profiling blocks it.
Document provider/deployer roles, human oversight, data governance and evidence.

What to assess

Which decision, assessment or access point the system supports.
Which data, scores or recommendations determine the output.
Who provides human oversight and which explanations are available.

Next step for this use case

Use the classifier for an initial route check or discuss the use case when the AI output may affect access, assessment, priority or rights.

Frequently asked questions

Short answers for classification, evidence and next steps under Annex III.

When does Emotion recognition fall under Point 1(c)?

AI systems for emotion recognition are high-risk (Annex III point 1(c)). In the workplace and in education, emotion recognition is prohibited, except for medical or safety reasons (Article 5(1)(f)). Assess whether the intended purpose falls within Point 1(c), then document whether Article 6(3), profiling and concrete rights impact change the classification.

Which practical examples should be assessed?

Examples to assess include: Emotion analysis of customers in service calls or at a service desk. and Inferring the mood or frustration of visitors from face or voice data, outside work and education.. The decisive point is not the tool name, but the function, output and consequences for natural persons.

What is the next step for this use case?

Use the classifier for an initial route check or discuss the use case when the AI output may affect access, assessment, priority, safety or rights within Biometrics.