Mindshare x amp - AI Voices vs. Real Voices

In the rapidly evolving landscape of audio technology, the debate between AI voices and human voices is more relevant than ever. Join us at SXSW 2025 for an insightful session that dives deep into groundbreaking research comparing the emotional, neurological, and perceptual impacts of AI and human voices on audiences.

A more detailed presentation will be presented day of and can be shared prior to SXSW.

1. Audio Context Effects

2. Background of Study

3. Background MINDSHARE TEMPLATE REVAMP 4 A Sample of the Questions Asked Included PURPOSE This study aimed to determine the different emotional, neurological, and perceptual reactions elicited by two separate voice sources: AI and humans. As AI voices increase in both abundance and quality, it is vital to understand how they are perceived by audiences on both an explicit and neurological level. Our findings will help us to understand what makes most effective brand voices worldwide and guide clients to the most successful AI and/or human voices for their brands with the most data-driven knowledge possible.    Is there a statistically significant difference between average response to AI voices and average response to human voices in this study?                                                    Do younger respondents (18-30) having a more positive response to the more familiar AI voices compared to the generic AI voices?                                             Do participants who claimed they used AI voices regularly have a more positive emotional response to the voices?                                                                          Is there a statistically significant difference in an ad’s impact on purchase intent when voiced by AI vs. when it was voiced by humans? Does this difference/similarity hold when accounting for participants ability to differentiate AI vs. Human voices?                                                                   And more….     

4. Method Lab-Based Research Approach A total N of 55 were recruited to participate in an in-lab experiment in a state of the art research labs based in Chicago, Austin, and New York. Participants were recruited as users of social media, podcasts, and radio, and randomly assigned to a rotation of audio clips. ​ All participants took place in the experiment in a simulated mock living room designed seated at a desk facing a computer equipped with headphones. Experimental software walked the participants through a series of audio clips which they would listen to and rate throughout the session. Neurometrics : Skin conductance and heart rate were collected as passive measures of physiological engagement and cognitive attention. ​ EEG : Electroencephalography was collected so that researchers could assess scalp-based neural activity as measures of approach and avoidance, as well as cognitive effort. ​ Survey : After exposure to each clip, participants were asked a few questions about that clip. A post-exposure survey was administered after all clips were completed. Measures