Youths' perceptions of AI chatbots as friends

The availability of conversational artificial intelligence (AI) is increasing. One new available chatbot is called My AI and was recently launched on the popular social media platform Snapchat, where the majority of Danish youth communicate daily. My AI is presented by Snapchat as a “new virtual friend” to all their young users. Although the chatbot is presented as a friend it does not automatically imply that Snapchat users also engage in friendship with the chatbot. This study investigates an important knowledge gap in the literature regarding how young Snapchat users perceive and relate to this new anthropomorphic social chatbot. Four focus group interviews and one paired interview with young Snapchat users were conducted to uncover young people’s perspectives on what is a good friend on Snapchat, and how they understand and perceive My AI as a friend. An additional five individual interviews were also conducted to gain more knowledge on the topic.

In the analysis of the data a theoretical framework of socialized affordances and friendship as contextual is used. The analysis establishes that young people use Snapchat primarily for personal contact with their friends; and specifically with friends whom they know offline. The analysis also reveals that a good friend on Snapchat is characterized by a reciprocity and engagement in contact, where you can trust the other with the personal snaps you send them. Furthermore, young Snapchat users recognize the socialized affordance potential in My AI as a friend, but choose not to actualize it. They perceive My AI to be creepy, unpersonal, lacking reciprocity and trustfulness and they relate to it in a critical, distancing manner. Their network context of already having a lot of offline friendships become part of their argument for not needing an AI friend. My AI becomes not only an unactualized affordance, but a downright deviant affordance, where the participants view a potential use of My AI as a friend as problematic.