AI in Human Resources: When Is the Line Crossed?

Especially when it comes to hiring, promotions, or even layoffs, a person must always have the final say—not a machine. Certain AI applications cross many boundaries in this area and therefore have no place in the sensitive field of human resources. These include systems that manipulate people without their knowledge or deliberately exploit psychological weaknesses. Emotion recognition in the workplace is also prohibited: Software may not use facial expressions, voice, or body language to infer stress, motivation, or suitability. AI-based social scoring systems are just as problematic. If email response times, location data, or personal information are aggregated into an overall score and this score influences personnel decisions, a company that allows this is operating in the red zone—an area subject to penalties under the EU AI Regulation.

Added to this is the risk of indirect discrimination: for example, if an algorithm views career breaks negatively, it can systematically disadvantage parents or older employees. Employers and HR departments need to keep this in mind as well. In such cases, clearly written guidelines are often helpful even before comprehensive training takes effect…