A new study has shown that job candidates change how they present themselves when they believe artificial intelligence is judging their suitability for a role. The findings, published in the journal PNAS, suggest that AI in recruitment could inadvertently narrow the range of qualities employers see in applicants, affecting hiring decisions across industries.
Researchers from Rotterdam School of Management and the University of St Gallen analysed how people adapt their behaviour during assessments when they know AI is involved. The study involved over 13,000 participants across 12 experiments. It revealed a clear trend: applicants tend to highlight their analytical abilities and downplay more intuitive or emotional qualities when they think they are being evaluated by AI rather than a person.
The research team pointed out that this shift stems from a common belief that AI values analytical skills above all else. Candidates expect algorithms to prioritise logic and data-driven traits, so they adjust their self-presentation to match what they think AI is programmed to prefer. This change occurs even when AI might be capable of recognising emotional intelligence and other less quantifiable attributes.
The findings matter because they highlight how AI does more than streamline recruitment processes. It also influences candidate behaviour, which could affect the fairness and effectiveness of hiring decisions. Employers could end up favouring people who adjust well to perceived AI preferences rather than those who are genuinely the best fit for a role.
In one part of the study, job applicants on platforms such as Upwork described themselves as significantly more analytical when told an AI would assess them. In other trials, participants were asked to rank personal qualities in order of importance. Those who believed AI would review their applications ranked analytical traits higher than those expecting a human assessor.
The study also tested ways to reduce this effect. When participants were encouraged to reflect on the possibility that AI could value intuition and emotion as well as analysis, they presented a more balanced picture of themselves. This suggests that clearer communication about AI’s capabilities could help candidates feel less pressure to alter their natural qualities.
The results come at a time when AI is increasingly used in hiring decisions, from initial application screening to interview analysis. As transparency regulations such as the European Union’s AI Act require companies to disclose their use of AI, more candidates are likely to become aware of these tools, potentially amplifying the behavioural shifts identified in this research.
The researchers warn that if organisations do not address this issue, AI could unintentionally reinforce biases or create new ones. They recommend that firms regularly review their AI tools and provide clear information about how assessments are made, helping to ensure that candidates feel able to present their true selves.
AI in recruitment does not just change the process, it changes how candidates behave and could reshape who gets hired.
