In this study we test how individuals react to the use of AI in different stages in the recruitment process. We do so by conducting an online experiment on Prolific with about 1500 participants. Therefore, we do not only randomly allocate participants to one out of four treatments, we also differentiate based on their highest educational level and work experience to conduct the experiment for two different versions; one job advertisement for high-skilled employees and one job advertisement that requires significantly less skills. Our main finding is that the use of AI reduces individuals' willingness to apply, for the high-skilled job even if the AI is used in the preselection phase only. Whereas the effect is much stronger for individuals with lower technological affinity, the effect also remains significant for those with a rather high affinity towards new technologies. For the low-skilled job, the negative effect appears only if the AI is also used in the interview stage of the recruitment process. The effect on participants' willingness to apply seems to be mediated by the perceived fairness of the process, which is perceived significantly less fair if an AI is involved, even in the treatment in which we explicitly highlight that the company policy ensures that the AI has been trained to be fair.