EU Bans Workplace Emotion AI, Microsoft Retired Azure Face Features in 2022

3 articles · Updated · Computerworld · May 15

The EU last year prohibited emotion-recognition AI in workplaces and schools, allowing only narrow medical or safety exceptions as regulators harden against employee-monitoring tools.
More companies are backing away because the systems often rest on weak science: a 2019 meta-analysis of 1,000-plus studies found facial movements alone cannot reliably reveal internal emotions.
A 2024 Finnish case study found workplace emotion tracking can undermine wellbeing, misread mental states, show racial bias, and expose identities despite claims of anonymous aggregation.
Microsoft had already dropped Azure Face emotion-recognition in June 2022, saying emotions lack scientific consensus and the tools raise privacy, stereotyping, discrimination, and unfair-denial risks.
That mix of legal limits, shaky accuracy, and surveillance concerns is pushing multinationals toward the stricter European standard, even as some safety-focused uses remain defensible.

Banned in Europe and scientifically flawed, why is the emotion AI market projected to triple?

As AI that reads emotions fails, is human empathy now the ultimate career superpower?