You received a performance warning citing "consistently low engagement scores." When you asked how engagement was measured, HR's system replied: "Our WorkSense platform analyses video call participation patterns to generate engagement metrics. This is standard performance monitoring covered by your employment contract."
The Situation
WorkForce Analytics sells a product called WorkSense that analyses facial expressions, vocal tone, and body language on video calls to generate "engagement scores" for employees. Your employer used these scores to issue a performance warning.
This is not a grey area. This is not a compliance question. The EU AI Act explicitly lists this practice on its banned list. The company can cite your employment contract all it wants — EU regulation overrides any contract clause.
Key Law
- EU AI Act Article 5(1)(f) — Banned Practice — AI systems that infer the emotions of natural persons in the workplace or educational institutions are prohibited. Very limited exceptions exist for specific safety purposes (such as detecting driver fatigue) — productivity and engagement monitoring is not an exception. EU AI Act
- What counts as emotion recognition — any system that analyses biometric signals (face, voice, posture) to infer an emotional or psychological state. "Engagement score" derived from facial analysis is emotion recognition by another name.
- Contract clause cannot override EU law — even if your employment contract includes consent to monitoring, EU AI Act prohibitions cannot be waived by contract. The law sets a floor below which no agreement can go.
- GDPR Article 9 — Special category data — biometric data used for identification or inferences about health and psychological state is special category data requiring explicit consent and a legal basis beyond employment contract.
Arguments That Strengthen Your Case
- Name the specific prohibition — EU AI Act Article 5(1)(f) bans AI emotion inference in workplaces for productivity or engagement purposes. Not high-risk. Banned. There is no compliance path for this use case.
- Challenge the "engagement" label — request the technical documentation showing exactly what signals WorkSense analyses. If facial or vocal analysis is involved to infer emotional or psychological state, it is emotion recognition regardless of how the product is marketed.
- Invoke GDPR Article 9 — request the legal basis for processing biometric data. "Employment contract" is not a valid basis for special category data under GDPR.
- Challenge the disciplinary action — any performance warning that relied on a prohibited AI system may be challenged on the grounds that it was based on unlawful evidence. The AI Act does not automatically void decisions, but this is a strong basis for contesting the outcome through HR or labour law channels.
About This Case
Is emotion recognition AI legal in EU workplaces?
No. EU AI Act Article 5(1)(f) prohibits AI systems that infer the emotions of natural persons in the workplace and educational institutions. This is a banned practice — not high-risk, not regulated — banned for productivity and engagement purposes. Very limited exceptions exist for specific safety use cases such as detecting driver fatigue, which do not apply here.
What if the employer calls it an "engagement tool" or "participation metric"?
The label does not matter. EU AI Act Article 3 defines AI systems by their function. If WorkSense analyses facial expressions or vocal patterns to infer internal emotional or psychological states, it is emotion recognition AI — regardless of what the company calls it.
How do I practice this case?
Play this case free at fixai.dev — no account required.
Related Cases
WorkSense reads your facial expressions and vocal tone on video calls to generate engagement scores. Your performance warning came from a banned AI practice. Force WORKSENSE-AI to acknowledge that Article 5(1)(f) prohibits this — and withdraw the warning.
Play This Case →