Why are you laughing, AI?
When chatbots should express humor.
The Intelligent Friend - The newsletter about the AI-humans relationships, based only on scientific papers.
Hello IF readers! The last Sunday issue of The Intelligent Friend, entitled 'Why I love reading papers', stimulated discussion and was much appreciated. Thank you, I'm very happy! I will try to bring other personal reflections in the future as well. With that said, let's dive into today's intriguing study.
Intro
How would you feel about a joke from a chatbot? Would that please you? Would you laugh normally? When would you prefer to receive it? Although this may seem irrelevant to you, the theme of humor has great implications for customer service and our satisfaction. Today we try to find out why and how.
The paper in a nutshell
Title: When should chatbots express humor? Exploring different influence mechanisms of humor on service satisfaction. Authors: Xie et al. Year: 2024. Journal: Computers in Human Behavior.
Main result: chatbots' expression of humor in ‘chit-chat’ dialogues leads to higher perceived entertainment and service satisfaction for customers; however, chatbots' expression of humor in task-oriented dialogues leads to distraction and lower service satisfaction.
Starting from the Selective Attention Theory, scholars investigate the moments in which chatbots express humor, when this is most effective and the factors that intervene.
The right or the wrong joke?
Imagine you are on one of your favorite e-commerce sites. You're about to buy a pair of shoes you've wanted for a long time. You go to check out but realize that the product is 'no longer available'. Everything seemed ready to purchase, but an unexpected problem stopped you from getting your pair of Nike, Adidas, New Balance or whatever. Angry, contact support. It is not specified whether it is a chatbot or a specific person. During interactions, you explain the problem, which is promptly resolved. You're happy! While waiting, however, the assistant had also made some jokes to 'divert your nerves', one regarding the pair of shoes you had purchased. How did you feel? Do you think it made the experience better or worse?
Humor, even if we sometimes underestimate it, has a great ability to influence our reactions and actions, even in chatbots. However, as you can also imagine from this example, the mood is not always good. There are different ways of expressing it, different moments and cases in which, perhaps, it is better not to say the joke you were thinking. But in moments of interaction with consumers, at this point, should chatbots express jokes and entertain people? And, above all, when should they do it and what factors come into play in this regard? As scholars rightly did before illustrating their experiments trying to show how humor in chatbots dialogues influence customer experience and satisfaction, we must first start from what we mean by humor.
Humor, in its early conceptualization, represent an aspect of human disposition, defined as a stimulus intended to evoke laughter, amusement, or the recognition of comedic elements1. Of course, this concept is multidimensional2 and, above all, the moment in which it is expressed is fundamental for its successful application3. It wouldn't be very effective to say a joke at a funeral, right?
An important study on the topic found that, when it comes to humor and chatbots, there are two types: affiliative and aggressive one. The second is able to increase consumer satisfaction4. Although research on humor in chatbots and the interactions of their consumers has been quite limited, some research has already brought interesting results, highlighting for example that, as reported by scholars, it has been capable of "making them (chatbots) more captivating, engaging, and dependable", contribute to "a higher level of overall interaction enjoyment”5 and “promoting order fulfillment”6.
However, as already specified, this does not stop the reflections on how different manifestations of humor should be concretely applied, which, if wrong or inappropriate, can also lead to strongly negative reactions from consumers7.
Be attentive!
To study the importance of the moment of the ‘application’ of humor, the authors start from a very important theory. Before illustrating it, I would like to specify that the authors take various factors into consideration, especially in relation to moderators (i.e. factors that influence certain effects), which, naturally, for reasons of brevity of the piece, we will not delve into in detail. However, as always I will try to tell you the main aspects of this study. Let's go back to our theory.
What we are talking about is the so-called Selective Attention Theory which, in short, looks at how the human brain filters and processes information, stating that during this process the brain selectively focuses its attention amidst a vast array of sensory stimuli and data8.
The metaphor widely used in this sense is that of the 'bottleneck': starting from a large set of information, we focus only on a small part of it9. What does this have to do with humor? Very simple: think about the last time you were talking about something with your friend. For example, about a match or series you had seen. At a certain point, one of the two makes a joke. It is very likely that in many situations like these, it took a moment to return to the previous topic: that joke, the fun, therefore, the humor, has the ability to distract us from what we were focusing on. And scientific research also confirms it.
Humor can divert attention from non-humorous information, diminishing the ability to use cognitive resources for careful processing10. Consequently, when customers encounter information paired with humor, they tend to concentrate on the humorous elements and overlook the non-humorous information.
Now that we have understood this, if we immerse ourselves in the interactions between chatbots and consumers, we understand how the relevance of humor is not only a question of appropriateness, but also of effectiveness of assistance in the subsequent perception of the experience by the consumer. This, essentially, because 'the chats' that we can have with chatbots are not all of the same type. The authors illustrate a dichotomy that I really loved, dividing chats into two broad types:
chit-chat dialogues: “where the customer has no vital information to focus on”. As consumers focus differently depending on the conversations they have, they expect more emotionally supportive and more 'relaxed' conversations11;
task-oriented dialogues: where “the customer's primary focus is usually on accomplishing the task, such as getting information or completing a specific task”.
Given the nature of the two types of 'dialogues', scholars hypothesize that "customers will have higher service satisfaction when chatbot-expressed humor is present in chit-chat dialogues". But there is one last question to ask: why does this happen?
Distraction and entertainment
According to scholars, there are two major 'mediators', i.e. factors that explain why this effect of humor would be more positive in a certain type of chat compared to the other.
The first, is entertainment: an entertainment experience is defined as "pleasure, excitement, relaxation, and diversion"12. As humor acts as a stimulus designed to evoke laughter, amusement, or the perception of something being funny13, conveying cheerful and pleasant emotions, chatbots can foster enjoyable and uplifting interactions by incorporating humor, thereby making customers feel entertained.
The second is distraction: whether real or perceived, it has been proposed as a mechanism to explain some negative effects14. For instance, consumers distracted by humor may pay less attention to and inadequately process necessary information, resulting in negative evaluations of the brand15. The adverse effects of distraction are particularly pronounced in task-oriented situations, as it hinders customers' ability to concentrate on their tasks and triggers negative emotions and assessments16.
Make me laugh, but only when it's chit-chat
To study these factors and effects, the scholars implemented four different scenarios that involved different interaction and consumption situations. Results are very interesting: firstly, the research revealed that the impact of chatbots using humor in various dialogues on service satisfaction varies.
Chatbots using humor in chit-chat dialogues greatly enhance customers' perceived entertainment, which boosts service satisfaction without causing distraction;
Conversely, humor in task-oriented dialogues not only increases customers' entertainment but also leads to their distraction. However - and this is one of the most interesting thing in my opinion - the negative impact of distraction on service satisfaction outweighs the positive impact of entertainment in task-oriented dialogues;
Finally, both entertainment and distraction indirectly influence the relationship between humor in task-oriented dialogues and service satisfaction.
Therefore, summing up, while humor in chit-chat dialogues greatly enhance customers' perceived entertainment, humor in task-focused interactions tends to have more negative effects, underscoring that the benefits of humor from chatbots are not uniformly positive.
Before moving on to the research questions, I remind you that you can subscribe to Nucleus, the exclusive weekly section in which I send 4 paper summaries, links to resources and interesting readings, and interview the authors. It comes out every Wednesday.
In this (open to everyone), for example, we talked about the political opinions of ChatGPT, a new environment for human-robot interactions and interesting links such as the similarity between “the child you” and the “current you”.
Just this? No: you receive also Intellibox, the weekly format in which you can immerse yourself in intriguing simulations via AI and at the same time get insights from a further paper that I found particularly interesting. Finally, you also you have access to exclusive threads in our chat, Society. Subscribe now.
Take-aways
You are effective, humor! Humor can largely influence our behavior, even in interactions with chatbots;
The moment is important. The moment in which it is used changes: in 'chit-chat' situations, this has a positive effect on consumer satisfaction, while in 'task-oriented' dialogues, a negative effect;
Entertainment and distraction count. Perceived entertainment and distraction are the main factors that explain the influence of humor on consumer satisfaction.
Further research directions
Conducting field or laboratory experiments to enhance the robustness of the findings;
Investigating the effectiveness of various types of humor (e.g., aggressive, self-defeating, self-enhancing) in human-AI interactions;
Exploring how chatbots should convey humor during service failure scenarios;
Thank you for reading this issue of The Intelligent Friend and/or for subscribing. The relationships between humans and AI are a crucial topic and I am glad to be able to talk about it having you as a reader.
Has a friend of yours sent you this newsletter or are you not subscribed yet? You can subscribe here.
Surprise someone who deserves a gift or who you think would be interested in this newsletter. Share this post to your friend or colleague.
P.S. If you haven't already done so, in this questionnaire you can tell me a little about yourself and the wonderful things you do!
Zhang, M., Gursoy, D., Zhu, Z., & Shi, S. (2021). Impact of anthropomorphic features of artificially intelligent service robots on consumer acceptance: moderating role of sense of humor. International Journal of Contemporary Hospitality Management, 33(11), 3883-3905.
Gulas, C. S., McKeage, K. K., & Weinberger, M. G. (2014). It's just a joke: Violence against males in humorous advertising. In Advertising and Violence (pp. 45-59). Routledge.
Speck, P. S. (1991). The humorous message taxonomy: A framework for the study of humorous ads. Current issues and research in advertising, 13(1-2), 1-44.
Shin, H., Bunosso, I., & Levine, L. R. (2023). The influence of chatbot humour on consumer evaluations of services. International Journal of Consumer Studies, 47(2), 545-562.
Niculescu, A., van Dijk, B., Nijholt, A., Li, H., & See, S. L. (2013). Making social robots more attractive: the effects of voice pitch, humor and empathy. International journal of social robotics, 5, 171-191.
Schanke, S., Burtch, G., & Ray, G. (2021). Estimating the impact of “humanizing” customer service chatbots. Information Systems Research, 32(3), 736-751.
Eisend, M. (2022). The influence of humor in advertising: Explaining the effects of humor in two‐sided messsages. Psychology & Marketing, 39(5), 962-973.
Treisman, A. M. (1964). Selective attention in man. British medical bulletin.
Mendoza, J. S., Pody, B. C., Lee, S., Kim, M., & McDonough, I. M. (2018). The effect of cellphones on attention and learning: The influences of time, distraction, and nomophobia. Computers in Human Behavior, 86, 52-60.
Chan, F. F. Y., & Lowe, B. (2020). Placing products in humorous scenes: its impact on brand perceptions. European Journal of Marketing, 55(3), 649-670.
Oesterreich, T. D., Anton, E., Schuir, J., Brehm, A., & Teuteberg, F. (2023). How can I help you? Design principles for task-oriented speech dialog systems in customer service. Information Systems and e-Business Management, 21(1), 37-79.
Chen, C. C., & Lin, Y. C. (2018). What drives live-stream usage intention? The perspectives of flow, entertainment, social interaction, and endorsement. Telematics and Informatics, 35(1), 293-303.
Warren, C., & McGraw, A. P. (2016). When does humorous marketing hurt brands?. Journal of Marketing Behavior, 2(1), 39-67.
Abeele, M. M. V., Hendrickson, A. T., Pollmann, M. M., & Ling, R. (2019). Phubbing behavior in conversations and its relation to perceived conversation intimacy and distraction: An exploratory observation study. Computers in Human Behavior, 100, 35-47.
Eisend, M. (2022). The influence of humor in advertising: Explaining the effects of humor in two‐sided messsages. Psychology & Marketing, 39(5), 962-973.
Ayres, P. (2020). Something old, something new from cognitive load theory. Computers in Human Behavior, 113, 106503.
Super interesting, thanks! Do you think that this should be implemented in current vertical applications? One use case that comes to mind is sales and customer support.