The wild world of AI: How animal species can be used in AI-based services
//

Integrating wild animals into AI-based services

As artificial intelligence (AI) evolves, animal representations are becoming prevalent. But how do consumers perceive them?

Which animal would you rather take running advice from, a bear or a cheetah? Which animal would you listen to if you were learning a new language: a chicken or an owl?

As ridiculous as these questions may seem, with the rate at which different platforms and services are adopting artificial intelligence (AI), it’s not unreasonable to assume you may soon have an online running coach with fur, claws, and spots. Especially as services increasingly adopt zoonotic AI. That is, AI is represented in animal form.

We can already have AI represented as an Otter transcribe meeting notes for us through Otter.ai. Other examples include GoCharlie.ai (dog) and Duolingo’s owl. Animal forms are being used as representations of AI, performing tasks traditionally completed by humans, such as providing customer service or teaching languages.

How might cultural interpretations influence the acceptance of different animal representations in AI-based services? As cultural nuances significantly influence perceptions of different animals, zoonotic AI adoption may also vary culturally. For instance, while a dog-bot may be well-suited for Western cultures, it may not be ideal for Middle Eastern nations.

Anshu Suri

Zoonotic AI and consumers

To understand the effects of zoonotic AI on consumer adoption, researchers from UCD Michael Smurfit Graduate Business School, HEC Montréal, and Nottingham University Business School China conducted a series of studies. These included assessing participant responses to various zoonotic and humanoid robotic AI representations, such as an AI bear copywriting assistant, an AI parrot, an AI tiger fitness assistant, and more.

Results from initial studies did confirm that consumers are less likely to choose an AI over a real-life human provider when the AI is represented by an animal compared to a humanoid robot. This is due to consumers finding it difficult to associate the animal with the task. For example, it’s difficult to imagine a bear supporting you with copywriting.

The wild world of AI: How animal species can be used in AI-based services
Credit. Midjourney

The concept of prototypicality

The authors also build into the concept of prototypicality to explain consumers’ enhanced difficulty in delegating a task to an animal-like AI rather than a humanoid robotic AI. Typically, humanoid robots represent AI. If asked to imagine an AI, you would most likely describe it as a human-looking robot. It would be very unlikely for a person to imagine an AI represented by an animal.

However, the authors show situations where the negative effect of using an animalistic AI could be mitigated. A follow-up study showed that consumers are equally accepting of AI in an animal form (as a humanoid robotic AI) if the task seems to fit the species, such as taking running advice from a cheetah. After all, cheetahs are one of the fastest animals on land!

Instinctively, humans often associate animals with human attributes. For example, many cultures around the world associate owls with wisdom and intelligence, which are important attributes when it comes to being a teacher. This could explain why an owl was chosen for Duolingo, an app used to learn languages.

Furthermore, in a final study, consumers were actually found to be more accepting of animal AI than humanoid AI if the activity, experience, or product was perceived to be fun and enjoyable. For example, wouldn’t you have more fun learning from an AI that a parrot represents than from a robotic AI? Simply put, when users are seeking efficiency, an animalistic AI may be a bad idea. However, when seeking an enjoyable experience, such as having fun while learning, an animalistic AI would be preferable to a robotic AI.

Practical implications

For those involved in digital product development, based on the findings of this research, questions should be asked regarding the effectiveness of using animal representations for delivering AI‐based services since consumers are less inclined to adopt these services when provided by zoonotic AI. Particularly given the substantial financial investments companies make in developing and marketing AI-based services.

However, if developers do choose to adopt an animal species to represent an AI product, consideration should be taken to ensure that the species chosen matches the nature of the service offered. Cultural interpretation also needs to be taken into account, as different cultures may perceive various animal species in different ways. For example, owls might be considered wise and intelligent in some cultures, a great representation of a learning service, but perceived as signs of misfortune in others.

Companies could also benefit from emphasising the entertaining aspects of the AI’s task to help ensure consumers are more likely to accept the AI animal. This can be seen in consumers’ adoration for Duolingo’s owl, as it helps them enjoy learning a new language, a task that could be laborious for some.

Conclusions

As a practical example, a few years ago, BarkBox, a company that specialises in dog toy subscription boxes, launched a dog bot (a chatbot using an AI represented by a dog to answer queries). It wouldn’t be a bad idea for BarkBox to understand the context of the query and use a humanoid or animalistic AI accordingly. For instance, a customer trying to understand different dog treats and toys, a seemingly enjoyable activity, could interact with a dog bot. However, a customer trying to return a product would ideally be served by a humanoid robotic AI, as this context seeks efficiency and functionality.

These findings advance the understanding of consumer–AI interactions in the context of zoonotic designs and provide valuable managerial insights into when and how firms should use animal representation for AI‐powered services. By examining the impact of robotic and zoonotic AI prototypes on consumer adoption, this research opens up possibilities and avenues to be further explored on a topic that has received limited attention.

🔬🧫🧪🔍🤓👩‍🔬🦠🔭📚

Journal reference

Poirier, S. M., Huang, B., Suri, A., & Sénécal, S. (2024). Beyond humans: Consumer reluctance to adopt zoonotic artificial intelligence. Psychology & Marketing41(2), 292-307. https://doi.org/10.1002/mar.21934

Dr. Anshu Suri is the UCD Garfield Weston Assistant Professor of Marketing at the UCD Michael Smurfit Graduate Business School. Dr. Suri's research focuses on understanding consumer responses in the face of brand and service failures, shedding light on critical aspects of consumer behaviour. Her scholarly pursuits also include consumer-technology interaction and the sharing economy, contributing to a deeper comprehension of contemporary market dynamics.