Classified examples · HR and employees · Po polsku

AI monitors and scores employees' productivity

Software tracks employees' activity (applications, keystrokes, calls) and uses AI to score productivity and flag low performers.

How we classify it

1. Annex III point 4(b) lists AI intended to monitor and evaluate the performance and behaviour of people in work relationships, and to allocate tasks based on individual behaviour or personal traits: high-risk.
2. If it infers emotions or stress from faces or voices, it is prohibited at work (Article 5(1)(f)).
3. Result: high-risk, with deployer duties from 2 December 2027; emotion inference is banned now.

What applies now

• GDPR and national employment law on monitoring apply now (in Poland, the Labour Code sets the rules for monitoring employees)
• No emotion or stress inference
• Human review of any consequence for an employee

What would change the answer

• From 2 December 2027: inform workers' representatives and the affected workers before use (Article 26(7)), trained human oversight, logs, explanation rights

Evidence to keep

• Classification record
• Impact assessment
• Information to staff and representatives

In the AI Compliance Kit: 04Risk Classification Procedure · 09AI Impact Assessment Template · 03AI System Inventory.

Dates and sources

• High-risk obligations for stand-alone systems (Annex III): 2 December 2027
• Prohibited AI practices, Article 5: 2 February 2025

Legal basis: AI Act art. 5(1)(f), 26(7), Annex III point 4(b); GDPR; national employment law.

Similar examples

• An AI tool screens and ranks job applications High-risk (Annex III)
• A video interview tool scores candidates' emotions Prohibited (Article 5)
• AI writes job ads, or decides who sees them Depends on the use