This work presents the development of semantic attributes for Acoustic Vehicle Alerting Systems (AVAS) in electric vehicles (EV), using psychoacoustic modeling and consumer preference analysis. Participants from different consumer groups were exposed to a series of AVAS sound samples and assessed them by means of semantic differentials. Preference classification across groups of consumers, together with correlation of the insights obtained with psychoacoustic parameters, is proposed in this paper to provide actionable guidelines for the optimal design of AVAS. The findings contribute to enhancing pedestrian safety and user experience while aligning with regulatory standards.