2025 is finally over, but other than as part of the liveblogs I haven't yet had a chance to round up our various presentations at conferences during the second half of the past year. We ended the year with the AANZCA conference on the Sunshine Coast, where I presented what was something of a labour of love: a look back on ten turbulent years of the #auspol hashtag on what used to be Twitter.
Through the efforts of a series of excellent data scientists in our QUT Digital Media Research Centre (DMRC) team (especially Brenda Moon, Felix Münch, Jan Tan, Laura Vodden), we managed to gather a nearly uninterrupted dataset of #auspol tweets from 2014 to the enxittification of the platform in 2023, and our interest in analysing this large dataset was especially in whether #auspol could be described at its core as a genuine online community. Here are the slides, and a book chapter covering this analysis is also forthcoming.
Axel Bruns and Anand Badola. “Ten Years of Uninterrupted Debate: The #auspol Hashtag Community, 2014-2023.” Paper presented at the AANZCA 2025 conference, Sunshine Coast, 26 Nov. 2025.
AI in/and Research
As for many others, too, one of the major themes of our analytical and methodological work throughout the year was generative AI, of course, and here I was inspired by Joanne Kuai's presentation at IAMCR 2024 to kick off some work of our own that analysed how AI chatbots respond to questioning that indicates an interest in conspiracy theories. The results are a mixed bag, with serious concerns especially over Grok – which will come as little surprise. We presented this research across a number of conferences, and there's plenty more to come in 2026.
Axel Bruns, Katherine M. FitzGerald, Michelle Riedlinger, Stephen Harrington, Timothy Graham, and Daniel Angus. “‘Just Asking Questions’: Doing Our Own Research on Conspiratorial Ideation by Generative AI Chatbots.” Paper presented at the IAMCR 2025 conference, Singapore, 16 July 2025; at the AoIR 2025 conference, Niterói, Rio de Janeiro, 17 Oct. 2025; and at the AANZCA 2025 conference, Sunshine Coast, 28 Nov. 2025.
In addition, of course we're also exploring constructive uses of generative AI and LLMs in specific research contexts. My colleague Laura Vodden is leading our effort to develop LLM-assisted frame analysis approaches, with some encouraging early outcomes. Still much more work to do, and fully featured frame analysis may well remain out of reach (and indeed, it turns out that human coders struggle massively with this task too) – but more limited applications may be possible. Laura's slides from the AoIR 2025 conference are here:
Laura Vodden, Katharina Esau, Axel Bruns, Tariq Choucair. “Extending Our Capabilities: Towards LLM-Assisted Frame Analysis of Australian Climate Movement News Coverage.” Paper presented at the AoIR 2025 conference, Niterói, Rio de Janeiro, 16 Oct. 2025.
LLMs have certainly proved useful already for large-scale content analysis: my colleague Carly Lubicz-Zaorski has successfully used them to study news reports and submissions to parliamentary enquiries in Australia. Also presented at AoIR 2025, her and our work on this clearly points to a number of discursive patterns that document the battlelines in the fight to drag Australian policy-making towards greater action on climate change:
Carly Lubicz-Zaorski, Katharina Esau, Laura Vodden, Tariq Choucair, Axel Bruns, Michelle Riedlinger, Ehsan Dehghan, and Samantha Vilkins. “Division and Delay in Australian Climate and Energy Discussions: An LLM-Assisted Analysis of Discourse Coalitions across News Reports and Parliamentary Submissions.” Paper presented at the AoIR 2025 conference, Niterói, Rio de Janeiro, 18 Oct. 2025.
Practice Mapping in Practice
Indeed, Carly's mapping of these discourse coalitions is also inspired by the practice mapping approach which we introduced in a Social Media + Society article last year. Having established the basic approach and its methodological considerations, during the second half of 2025 we began to apply this approach to a number of obvious and not-quite-so-obvious contexts, and to extend it further into a number of interesting directions. I also presented another overview of practice mapping at ZeMKI's 20th anniversary conference in Bremen:
Axel Bruns, Katharina Esau, Kateryna Kasianenko, Tariq Choucair, and Vish Padinjaredath Suresh. “Diagnosing Destructive Polarisation in Public Discourse: The Practice Mapping Framework.” Paper presented at the ZeMKI 20th Anniversary Conference, Bremen, 23 Oct. 2025.
For a paper presented by my colleague Sebastian Svegaard at AoIR 2025, we applied this approach to the various Taylor Swift fandom (and anti-fandom) subreddits on Reddit. This particularly showcases how practice mapping enables us to conduct network analyses of platforms that (because of their threaded nature) don't usually lend themselves to the extraction of network data:
Samantha Vilkins, Axel Bruns, Sebastian F.K. Svegaard. “Mapping Fandom Ruptures: A Case Study of Taylor Swift Fandom Practices on Reddit.” Paper presented at the AoIR 2025 conference, Niterói, Rio de Janeiro, 16 Oct. 2025.
We took this even further for an AoIR 2025 paper that reconsidered how we could still analyse election campaigns in a social media environment that is now highly diverse and disjointed, and where data on public activity on these platforms are increasingly difficult to access or to work with. I'm particularly pleased that, as a proof of concept, we managed to conduct at least some basic practice mapping even within the severely limited and cumbersome walled-garden environment of the Meta Content Liibrary – and our application of practice mapping approaches to election-related news content also turned out to be surprisingly productive:
Axel Bruns, Samantha Vilkins, Katherine M. FitzGerald, Tariq Choucair, Daniel Angus, Caroline Gardam, Kunal Chand, Laura Vodden, Klaus Groebner, Katharina Esau, Carly Lubicz-Zaorski, and Ehsan Dehghan. “Researching Cross-Platform Campaigning in the 2025 Australian Federal Election.” Paper presented at the AoIR 2025 conference, Niterói, Rio de Janeiro, 18 Oct. 2025.
Still very much work in progress, and presented at such at a number of conferences over the year, is our large-scale analysis of climate-related discussions on Facebook in Australia and Brazil. In particular, here we're hoping eventually to extend the practice mapping approach in a temporal dimension: to generate a dynamic practice map that evolves with the passage of time. Representative of a number of related presentations, here is the paper on this that I presented at AoIR 2025:
Axel Bruns, Carly Lubicz-Zaorski, Tariq Choucair, Laura Vodden, and Ehsan Dehghan. “Tracking Shifts in Discursive Alliances: A Longitudinal Analysis of Australian Climate Change Discourses on Facebook through Practice Mapping.” Paper presented at the AoIR 2025 conference, Niterói, Rio de Janeiro, 18 Oct. 2025.
Understanding Search Engines
And last but certainly not least, in the ARC Centre of Excellence for Automated Decision-Making and Society (ADM+S) we are now also seriously ramping up our research into search engine recommendations, including both organic search results and AI-generated information summaries. Here too our approach is partly inspired by practice mapping techniques, and I presented a first analysis of recommendation diversity patterns for related search queries at the SEASON 2025 conference in Hamburg:
Axel Bruns, Daniel Angus, Ashwin Nagappa, Kateryna Kasianenko, Abdul Obeid, Shir Weinbrand, and Brett Tweedie. “Assessing Recommendation Diversity in Search Results: Approaches Using Data Donations and Artificial Personas.” Paper presented at the SEASON 2025 conference, Hamburg, 25 Sep. 2025.
My colleague Kateryna Kasianenko then presented a further update on this work at the AANZCA 2025 conference in November, too – our focus here was especially on how search engines respond to queries about conspiracy theories:
Kateryna Kasianenko, Caroline Gardam, Katherine M. FitzGerald, Ashwin Nagappa, Daniel Angus, Shir Weinbrand, Samantha Vilkins, Axel Bruns, Abdul Karim Obeid. “Searching for the Truth? Search Engine Responses to Conspiratorial Search Practices.” Paper presented at the AANZCA 2025 conference, Sunshine Coast, 26 Nov. 2025.
And at the JERAA conference in Brisbane the following week, she also presented a similar study that focussed instead on the results Google Search returns for queries about Russia's war on Ukraine
Kateryna Kasianenko, Ashwin Nagappa, Silvia Montaña-Niño, Michelle Riedlinger, Ned Watt, Anand Badola, Axel Bruns, and Daniel Angus. “Facts and Fabrication in Search Experience: Russia’s War on Ukraine and Google's Role in Gatekeeping Fact-Checked Information.” Paper presented at the JERAA 2025 conference, Brisbane, 2 Dec. 2025.
We're planning to conduct similar case studies, and to advance our analysis of the results produced by Google Search, on a wide range of further topics in the new year.
More to Come...
And that's it for the last few presentations in 2025. Much more to come in 2026, of course, starting with our DMRC Summer School in early February, a keynote at the Social Media Access Days in Frankfurt, and continuing with a visiting Mercator Fellowship at ZeMKI in Bremen in March to June.
But we've also managed to get several new journal articles and book chapters over the line during these last few months, including one final article that came out on Christmas Eve. More on those in another post soon!
