MELLAN NYTTA OCH MOTARBETAD MÄNSKLIGHET. En kvalitativ intervjustudie av målgruppen millennials användning och upplevelse av AIverktyget ChatGPT i vardagen

Sammanfattning

AI-chatbots such as ChatGPT has rapidly become an integrated part of people’s everyday lives and the number of active users is increasing both globally and in a Swedish context. AIchatbots are different from previous media technologies in their interactive design and humanlike communication. While AI-chatbots are known for facilitating many tasks in work life, studies and leisure activities, questions are being raised about AI-chatbots’ potential to replace human social interactions and creativity, and experts are warning about blindly trusting AI-chatbots in providing information and making decisions. Research has defined millennials as a generation expected to be highly affected by the development of AI-chatbots while also having a role in shaping these technologies. The purpose of this study is to examine millennials’ uses and experiences of ChatGPT in their everyday lives. Through the theoretical lens of Uses and Gratifications the study attempts to answer the first research question “What gratifications motivates millennials use of ChatGPT in their everyday lives?”. The study also moves beyond the field of media and communication by applying theories such as Computers Are Social Actors (CASA) and The Uncanny Valley which originate from the research fields of Human-Computer-Interaction (HCI) and Human-AIInteraction (HAII). Through this approach, the study seeks to understand the experiences of ChatGPT in relation to its humanlike and interactive properties and answer the second research question; “How do millennials perceive and manage the boundary between human and machine in their interaction with ChatGPT?”. Through qualitative semi structured interviews, this study aims to address a research gap in a Swedish context with a focus on the everyday uses and experiences of ChatGPT. Most previous studies consist of quantitative methods, focusing on specific life contexts with participants mainly from Asia or the USA. The results show that millennials primarily use ChatGPT for practical tasks motivated by utilitarian gratifications with the specific gratifications time saving, processing existing material, information seeking and creativity. Technological gratifications arising from ChatGPT’s specific affordances are also considered important with gratifications such as convenience, user control, user profiling and novelty. Hedonic gratifications are seen as a less central use identifying the gratification exploring leisure interests. Social gratifications are found to be the least influential with participants taking a clear stance toward ChatGPT being used for social companionship. One social gratification was discovered: interpreting and simulating social interactions. The findings support previous research identifying utilitarian gratifications as the most important, while attributing somewhat greater importance to technological gratifications than previously reported and less significance to hedonic and social gratifications. The findings also largely support previous research showing that people treat and perceive media technologies according to social rules like politeness and sometimes apply human categories to them such as gender, ethnicity and personality with ChatGP often perceived to be male, white and have a submissive, nice and pleasing personality that could encourage dangerous decisions. The study strongly contradicts previous research claiming that sharing personal information to ChatGPT can result in the same kind of emotional support as to another human. In line with previous research this study found that people generally experience ChatGPT as strongly humanlike which sometimes could result in feelings of creepiness connected to The Uncanny Valley theory, especially regarding false information from ChatGPT. Unfulfilled expectations of humanlike behavior were, contradictory to previous research, not connected to feelings of creepiness but rather to amusement. The study showed a clear preference for a robotlike and mechanical communication style from ChatGPT which aligns with previous research on how people prefer to interact with robots. All this indicates a strong separation between human and machine in the interaction with ChatGPT. While the millennials in this study express a largely unproblematic use of ChatGPT in regard to themselves they raise concerns about more vulnerable groups and societal risks such as false information, collection of data, environmental impacts and a potential decline in personal cognitive and creative abilities.