TY - ABST
T1 - Making sense of audiences in a time of generative AI
AU - Heuser, Morten
AU - Vulpius, Julie
AU - Mahnke, Martina Skrubbeltrang
PY - 2024/6/19
Y1 - 2024/6/19
N2 - With the launch of ChatGPT 3.5 in October 2022, audiences everywhere gained instant, free access to an AI interface that provide an unprecedented offer of open-ended applications, intuitive use, conversational output and human-like interaction. Simultaneously, its creator, OpenAI, invited the public to take part in an explorative and configurative journey and contribute to the shaping of ChatGPT by experimenting and coming up with uses and applications (Spennemann 2023). By extension, this invitation also targeted domain-specific communities, ranging from health, business to economy and research.For audience research, practitioners are using ChatGPT for audience analysis and to gain new or faster insights when identifying audience segments, defining target audiences or personalizing audience content (medium.com). In research, the launch has prompted widespread experimentation and reflection across fields. Experimentation has yielded mixed results in terms of AI’s ability for qualitative analysis (Siiman et al. 2023, Morgan 2023), creative writing (Ippolito et al. 2023) and as a didactic tool (Kim & Kim 2022). Others have experimented to see if audiences can distinguish AI generated content from text written by humans (Gunser 2021). It has been described as a ‘novel tool’ and used to investigate cultural artifacts and social biases in storytelling both by humans and generative AI (Begus 2023).Taking cue from this background, we synthesize our recent audience research experiments using ChatGPT as a research tool, probe, didactic tool and object of research. Our experiments with audiences include using ChatGPT as:- an analytical partner in a thematic analysis of audiences’ adoption and sensemaking of ChatGPT.- an analytical tool to quickly transform non-literary social media data into standard forms, and to sort and categorize social media data about audiences- a novel research tool, inscribing it as a social actor and facilitator in a study experiment of Dungeons & Dragons player interaction to explore human/AI interaction and collaborative storytelling- to generate AI versions of ‘organic’ content (a press release, a news article and a PR text) in a reception study with 40 university students to explore if they could tell AI generated content from non-generated and to elicit their reception and reading strategies of AI generated content- a didactic tool with communication students to explore how they make sense of quality in ChatGPT outputs in the production of communication productsThe goal is to stimulate creative and critical discussion of the application of ChatGPT and related generative AIs in relation to audience research.
AB - With the launch of ChatGPT 3.5 in October 2022, audiences everywhere gained instant, free access to an AI interface that provide an unprecedented offer of open-ended applications, intuitive use, conversational output and human-like interaction. Simultaneously, its creator, OpenAI, invited the public to take part in an explorative and configurative journey and contribute to the shaping of ChatGPT by experimenting and coming up with uses and applications (Spennemann 2023). By extension, this invitation also targeted domain-specific communities, ranging from health, business to economy and research.For audience research, practitioners are using ChatGPT for audience analysis and to gain new or faster insights when identifying audience segments, defining target audiences or personalizing audience content (medium.com). In research, the launch has prompted widespread experimentation and reflection across fields. Experimentation has yielded mixed results in terms of AI’s ability for qualitative analysis (Siiman et al. 2023, Morgan 2023), creative writing (Ippolito et al. 2023) and as a didactic tool (Kim & Kim 2022). Others have experimented to see if audiences can distinguish AI generated content from text written by humans (Gunser 2021). It has been described as a ‘novel tool’ and used to investigate cultural artifacts and social biases in storytelling both by humans and generative AI (Begus 2023).Taking cue from this background, we synthesize our recent audience research experiments using ChatGPT as a research tool, probe, didactic tool and object of research. Our experiments with audiences include using ChatGPT as:- an analytical partner in a thematic analysis of audiences’ adoption and sensemaking of ChatGPT.- an analytical tool to quickly transform non-literary social media data into standard forms, and to sort and categorize social media data about audiences- a novel research tool, inscribing it as a social actor and facilitator in a study experiment of Dungeons & Dragons player interaction to explore human/AI interaction and collaborative storytelling- to generate AI versions of ‘organic’ content (a press release, a news article and a PR text) in a reception study with 40 university students to explore if they could tell AI generated content from non-generated and to elicit their reception and reading strategies of AI generated content- a didactic tool with communication students to explore how they make sense of quality in ChatGPT outputs in the production of communication productsThe goal is to stimulate creative and critical discussion of the application of ChatGPT and related generative AIs in relation to audience research.
M3 - Conference abstract for conference
T2 - Reviving Qualitative Audience Research for the Streaming Era
Y2 - 19 June 2024 through 19 June 2024
ER -