Critical wayfinding: How Users Engage with and Make Sense of GenAI conversations

This paper explores how users engage with and make sense of generative AI (GenAI). Positioning them as ‘GenAI audiences’, the study explores the cues and strategies used to interpret and negotiate meaning in the context of human-machine communication. The rapid adoption of ChatGPT has reignited critical questions regarding how humans ascribe agency and relational-ity to machines (Guzman & Lewis, 2020). The study takes its starting point in Lagerkvist’s (2024) question whether “somebody is at home” in the machine, or, in other words, why users may in-teract with GenAI as if it possessed agency or presence.

Current research has been emphasizing (infra-)structural power (van Dijck et al., 2018). How-ever, this perspective has been critiqued for overlooking and downplaying the lived user expe-rience in the context of everyday life (Mahnke et al., 2024). Algorithms and intelligent machines have been described as mirrors reflecting back user input, albeit in a distorted or curated way (Vallor, 2024) and further studies have shown that users are not just passive content consum-ers but actively (co-)construct meaning with GenAI (Mahnke & Bagger, 2024). Less attention has been paid to the complex navigational processes users need to engage in when communi-cating with GenAI. This study contributes to audience research with an empirical examination of how GenAI audiences create meaning in GenAI communication, and, furthermore, how GenAI’s technical affordances shape these processes.

Empirically, the study draws on data collected from Danish master students in communication studies, including 16 chat protocols, 18 reflection papers and five group discussions. The pre-liminary thematic analysis (Braun & Clarke, 2021) resulted in three key themes: 1) The dual na-ture of GenAI, 2) GenAI building human connections, and 3) Navigating ethical concerns. Users generally recognized ChatGPT’s potential as a conversation partner, however, they also re-mained critical towards limitations and known inaccuracies. They further acknowledged GenAI’s attempt to build human(-like) relationships and provide emotional support. However, they remained distanced, questioning GenAI’s authenticity and whether it could potentially replace humans.

Overall, the analysis suggests that users tend to position themselves as observers rather than co-creators when engaging with and making sense of ChatGPT. Especially when using GenAI as a dialogue partner, they aim to maintain a critical distance. Although this may reflect limita-tions inherent in the methodological setup, it still shows that users are actively positioning themselves in human-machine communication. We therefore propose the concept of “critical wayfinding” as a conceptual lens to capture and highlight the socio-technical aspects of mean-ing-making, showing that users not only make sense with the machine but also of the machine. This allows for a shift of perspective towards viewing GenAI as a dynamic interlocutor, where users need to find their way through interpretative cues, personal strategies, and epistemic boundaries. Extending this further, future research should explore how users co-construct dual meanings – both with and about GenAI – especially in the context of everyday life. This will con-tribute towards a more nuanced understanding of human-machine communication and inform ongoing discussions regarding user and machine agency.