"It's not my Fault!"

摘要

We investigated the effects of the deceptive behaviour ofa robot, hypothesising that a lying robot would be per-ceived as more intelligent and human-like, but less trust-worthy than a non-lying robot. The participants engaged ina collaborative task with the non-lying and lying humanoidrobot NAO. Apart from subjective responses, a more objec-tive measure of trust was provided by the trust game. Ourresults confirmed that the lying robot was perceived as lesstrustworthy. However, we have found no indication of theincreased intelligence or human-like perception of the robot.Instead the robot was perceived as less friendly, kind andresponsible. The results of trust game were aligned withthe results obtained via subjective responses showing thepotential of this indirect trust measure in the human-robotinteraction studies.