close

Research - 24.06.2025 - 13:45 

Artificial intelligence assessment changes human behaviour

Artificial intelligence is being used to assess people. For example, in the recruitment process, university admissions or the provision of public services. A study by the University of St.Gallen shows its affect on people.

Artificial intelligence (AI) is increasingly replacing human decision-makers – whether in job applications or university admissions. This has consequences. Researchers at the Universities of St.Gallen and Rotterdam show in a new study entitled “AI assessment changes human behaviour” that people behave systematically differently when assessed by AI.

AI assessment effect

In twelve studies involving over 13,000 participants, the team demonstrates a so-called “AI assessment effect”. Simply knowing that they are being assessed by AI changes how people present themselves – with potential consequences for selection decisions. The study was published in the prestigious multidisciplinary journal Proceedings of the National Academy of Sciences.  

Key findings of the study:

  • AI not only changes selection processes – it also changes the behaviour of applicants
    AI changes selection processes by automating and standardising decisions and promising to make them more efficient and objective. At the same time, AI also influences the behaviour of applicants – the mere knowledge that they are being evaluated by AI leads people to tailor their self-presentation to what they believe is appropriate for AI.
     
  • Changed thinking style: analytics instead of intuition
    When assessed by AI, people increasingly emphasise analytical, systematic and rule-based thinking. Emotional intelligence, creative abilities and intuitive qualities, on the other hand, take a back seat, even though they are just as important for many roles and are becoming increasingly important.
     
  • Lay understanding of AI among applicants
    The shift to analytical thinking is based on a widespread assumption: people believe that AI favours analytical characteristics – regardless of whether this is actually the case. This lay understanding actively influences the behaviour of applicants.   · Bias can lead to a loss of diversity If people behave differently under AI assessment, this can lead to companies no longer assessing the “real” person, but rather their strategically adapted self-presentation. This bias can unintentionally result in a loss of diversity in thinking styles and personality.
     
  • Identify the AI assessment effect in assessment practices
    In the long term, the question arises as to whether adjustments to the criteria preferred by AI influence the expression of individuality, intuition or emotional intelligence. Companies should take the AI assessment effect into account to ensure that they receive authentic responses from applicants.


Artificial intelligence is not only changing who makes decisions – it is also changing how people behave in the selection process. The study shows that simply knowing they are being assessed by AI causes people to adapt their self-presentation and emphasise analytical characteristics more strongly. This can lead organisations to unintentionally favour applicants who present themselves in a particularly AI-friendly manner – not necessarily those who are best suited to the job profile. With the growing use of such systems, a fundamental question arises as to how human behaviour will change in a world where machines are increasingly becoming evaluators. Organisations and society are called upon to consciously shape this development in order to ensure diversity, authenticity and fairness in the digital age.

“We become what we believe AI wants to see: analytical, rule-compliant, predictable. In the process, our humanity is lost.”
Prof. Dr. Emanuel de Bellis, study author

Study method

A total of twelve studies involving 13,342 participants examined how people behave in job application situations when they are evaluated either by AI or by a human being. The studies included both real-life job applications and controlled online experiments. The participants were analysed in terms of how they described and presented themselves under different evaluation conditions. The recurring finding was that: The assumption of AI evaluation consistently and systematically changes human behaviour.  


The study was published in the journal Proceedings of the National Academy of Sciences and is available for free download. 


Image: Adobe Stock / Kromstar Studios

north