(Gambino et al., 2020)

The reason is probably related to the "Computers Are Social Actors" (CASA) paradigm. This theory suggests that humans often interact with computers as if they were social beings. However, the perceived lack of genuine consciousness, feelings, and social judgment in AI can reduce the pressure to maintain a socially desirable persona. Candidates may feel that the AI is a less judgmental evaluator, leading to more straightforward and less embellished responses.