DAVIS, CALIFORNIA - The use of artificial intelligence (AI), robots, and smart objects has recently risen, as most people know, but far fewer are aware of objects designed to resemble various animals - a practice known as zoonotic design. 

The emerging trend of using such zoonotic prototypes for AI technology in particular can be observed across different domains - from chatbots like BarkBox’s dog bot to services like Otter.ai’s otter. Yet most managers and marketers have little practical guidance on when and why these zoonotic AI embodiments should be used. 

If one is asked what kind of image comes to one’s mind when thinking of AI, a common response would probably be a humanoid robot. This association of ‘AI-as-robots’ is likely influenced by the prevalence of such portrayals in movies, media, and interactions with AI-powered entities - including chatbots and service robots - which often exhibit humanistic characteristics such as a human-like appearance, name, or even gender. If a zoonotic design is used instead, what impact does this have on consumer adoption of AI?

Through seven studies I undertook with my colleagues from HEC Montréal and Nottingham University Business School China, we explored whether the use of zoonotic AI embodiments positively or negatively affected consumer adoption of such services. To do this, we built on the concept of prototypicality - a measure of how representative an object is of a category. As noted earlier, a humanoid robotic design is a prototypical representation of AI in consumers’ minds.

Our research findings suggest zoonotic AI designs sometimes hinder consumer adoption when compared to more prototypical robotic embodiments. Distinguishing between robotic avatars - which tend to be humanoid in appearance - and zoonotic ones, we argue that zoonotic AI designs have this impact on consumer adoption due to an increased cognitive difficulty in associating the task