EU’s New AI Act and Its Implications for Emotion Recognition Systems
The European Union’s recent legislative measure, the AI Act, introduces significant restrictions on the use of emotion recognition systems within workplaces. This act aims to address privacy concerns and potential biases associated with AI-driven emotion detection technologies.
Scope of the Act
The AI Act specifically prohibits AI systems that infer emotions through various means, including keystrokes, body postures, and movements, as these methods rely on biometric data.
The guidelines are crafted to ensure a “consistent, effective, and uniform” application of the AI Act throughout the European Union. Although these guidelines provide insight into how the Commission interprets the prohibitions, they are non-binding, with final interpretations resting with the Court of Justice of the European Union (CJEU).
Emotion Recognition AI Systems
The restrictions extend to the use of emotion recognition AI systems during critical phases such as recruitment and probationary periods. Notable prohibitions include:
- AI systems monitoring the emotional tone in hybrid work teams, which identify and infer emotions from voice and imagery during video calls.
- Supermarkets using cameras to gauge employees’ emotions, such as happiness.
- Call centers employing webcams and voice recognition systems to track emotions.
However, there are exceptions for the deployment of these tools for personal training purposes, provided that data privacy is maintained, and the information does not influence employee assessments or promotions.
Addressing Privacy Risks
The primary aim of the AI Act’s restrictions is to mitigate the privacy risks associated with emotion recognition technologies. Such systems raise ethical concerns due to the diverse ways individuals express emotions across different cultures and contexts. By prohibiting these practices, the legislation seeks to uphold workers’ dignity and prevent discriminatory practices.
Nevertheless, exceptions remain for AI systems employed to monitor emotional states for medical or safety-related purposes. For example, in high-risk environments like factories or construction sites, AI might be utilized to detect signs of stress or burnout, provided that the technology directly relates to employee well-being and safety.
Conclusion
Enacted in August 2024, the EU’s Artificial Intelligence Act will be fully applicable by August 2, 2026. The ambition behind this legislative effort is to provide legal certainty for those deploying AI systems in the European market, as well as for market surveillance authorities. Although the guidelines are not legally binding, they represent a significant step toward regulating the use of AI in workplaces, particularly concerning emotional analysis and biometric data.
