Computer-Generated Speaker Charisma and Its Effects on Human Actions in a Car-Navigation System Experiment - or How Steve Jobs’ Tone of Voice Can Take You Anywhere

Abstract

A particularly persuasive (charismatic) tone of voice has a far reaching influence on people's opinions and actions. However, does this also apply if the charismatic tone of voice is produced by a computer, and if this computer asks people to act against better knowledge? Addressing these questions, an experiment was set up in which 30 locals of Sonderborg/DK were asked to conduct a test drive with a car from the marina to the university campus of the city. The test drive was conducted on the pretext of assessing a newly developed retrofit car navigation system that provides voice instructions only. The locals did not know that the system was only a remote-controlled mockup that, moreover, started giving its driver wrong instructions after about half of the trip. The instructions got successively worse, until the only option to get to the university campus was to make a complete U-turn. We measured the point at which the drivers aborted the test drive in two conditions. In the first one, the system spoke with the more charismatic tone of voice of Steve Jobs. In the second one, it spoke with the less charismatic tone of voice of Mark Zuckerberg. Results show that drivers followed the navigation system's increasingly worsening instructions significantly longer in the Steve Jobs condition, and that the system received higher quality, trustworthiness, and purchase ratings if it spoke with Steve Jobs' tone of voice. Results are discussed in terms of persuasive technology and implications for charisma/leadership analyses and training.

Myth busted? Challenging what we think we know about charismatic speech

“…Another technology-driven way of investigating charisma and its impact on listeners is using robots and talking machines. Works of Fischer (2018) or Niebuhr and Michalsky (2019) show that listeners interpret the acoustic charisma cues in synthetic voices in the same way as for human speakers. But, unlike for human speakers, the acoustic output of talking machines can be precisely controlled and exactly replicated.…”

“…Chen et al (2014), who conducted a similar analysis, arrived at the conclusion that auditory cues outweigh visual cues as a predictor for charisma. Lastly, the studies done by Fischer (2018) as well as Niebuhr and Michalsky (2019) show that computer voices possess charismatic influence even if the lexical material is identical, visual cues are absent, and only acoustic cues serve to signal charisma.…”

“…Ning (2019) shows that participants in a brainstorming workshop rate themselves as more intelligent, unconventional, and capable, when the workshop is given by a more charismatic moderator. Lastly, the study by Niebuhr and Michalsky (2019) shows that listeners even project the charismatic features of trustworthiness and competence to a computer system when receiving instructions with a charismatic tone of voice.…”

Advancing higher-education practice by analyzing and training students' vocal charisma: Evidence from a Danish field study

“…A second reason is that charisma is not just about persuasion. In particular, the non-verbal, voice-related features of speaker charisma, which are also among the most powerful ones (Chen et al 2014), express competence, self-confidence and passion (Michalsky & Niebuhr 2019). This, in turn, creates trust, motivation, and commitment on the part of the listener, even if the speaker is not a human.…”

“…This, in turn, creates trust, motivation, and commitment on the part of the listener, even if the speaker is not a human. Talking robots or cars whose speech synthesis was enhanced for vocal charisma can make human interaction partners eat healthier food and take detours in their own city (Fischer et al 2019;Niebuhr & Michalsky 2019). A stronger speaker charisma among human team members can increase the cohesion, interactiveness, and output of a team as well as its creativity (Stashevsky et al 2006;Amabile et al 2004).…”