Artificial Intelligence (AI) is an increasingly prevalent field that influences a number of other areas. It generates a multitudeof reactions among the general population, particularly anxiety, which impacts on the development, deployment, andregulation of AI. Nevertheless, experimental data on public perceptions toward AI are generally lacking. To fill thisgap, this paper presents a large-scale experiment conducted on the influence of the terms used to describe AI on peoplesperception. In a preliminary study (705 participants), words related to AI were extracted. In a second experiment (552participants), the impact of these terminologies on anxiety toward AI was explored. An unprecedented effect of wordingand a positive bias of perception toward computers was revealed by these experiments, compared to robots and newtechnologies. This research paves the way for future studies on the effect of words on perception in the field of AI andnew technologies.