https://doi.org/10.25547/RRQA-EN62
Lisez-le en français
This observation was written by Alan Colin-Arce with thanks to CAPOS Partner Michael Falk for his comments and review.
At a Glance
| Topic / Titre | Recap of the 7th annual gathering of the Canadian Australian Partnership for Open Scholarship |
| Key Participants / Créateur | Canadian Australian Partnership for Open Scholarship |
| Date / Période | 2025 |
| Keywords / Mots-clés | CAPOS, open social scholarship / approches sociales des savoirs ouverts |
Summary
The 7th annual gathering of the Canadian Australian Partnership for Open Scholarship (CAPOS), held on 2-3 December 2025 at the Australian National University in Canberra, brought together a vibrant group of scholars, policymakers, industry experts, and advocates to discuss recent developments in artificial intelligence and open scholarship. This year’s theme “Re-Defining Open Social Scholarship in an Age of Generative ‘Intelligence’” provided reflection points into how artificial intelligence is changing the open scholarship landscape for Canadian and Australian researchers and stakeholders.
Day 1
The first day of the conference began with a welcome and acknowledgement of country by Tully Barnett and Ray Siemens, as well as a shoutout to the winner of the 2025 Open Scholarship Award: Nick Thierberger and the PARADISEC team for their work on the project Pacific and Regional Archive for Digital Sources in Endangered Cultures (PARADISEC). This project has digitized over 18,500 hours of audio recordings and 4,000 hours of video recordings of materials in 1,400 languages across the world, particularly in the Pacific Region.
Following the welcome, Tyne Daile Sumner gave the opening keynote titled Creative Critique: Art, Open Scholarship & Generative AI. In the talk, Sumner discussed projects such as Art in the Cage of Digital Reproduction to exemplify how AI art can do a Creative critique of open data. The keynote concluded with some further questions for reflection on how to assess provenance in AI art, how to protect artists who expose unethical data practices, and what the implications are of generative AI in how art is critiqued.
The first day also featured two plenary panels. In the first panel, leaders of Australian humanities and library associations shared their provocations regarding the role of humanities and university libraries in the age of AI, such as Frank Bongiorno (Council for the Humanities, Arts and Social Sciences), Kylie Brass (Australian Academy of the Humanities), Tom Foley (Council of Australian University Librarians), and Chris Hay (AusStage). The presenters highlighted that AI is changing the processes of knowledge production and warned of the risks of reproducing the biases and gaps found in AI when using these tools to identify gaps in library collections.
In the second plenary panel, Amanda Lawrence described how Wikipedia is an open source infrastructure that shares digital public goods using a collaborative model based on the commons. Ian McCrabb discussed how Systemik Solutions is integrating AI into their workflow when supporting digital humanities projects. Michael Falk the development of a fully open large language model (LLM) and questioned whether these models can be truly open given the high technical and infrastructural barriers to develop LLMs.
The three presentation sessions also discussed the intersections of AI and open scholarship in multiple ways. Roxanne Missingham addressed the history of Creative Commons licenses and the challenges they face due to AI bots crawling large amounts of data from the web (as discussed in a previous OSPO post). Jiadai Xia described her study on the use of AI tools by PhD students whose first language is not English, and Leah Henrickson presented on the use of data donations of social media data for research and teaching about the extent of datafication and surveillance in social media platforms.
The second presentation session covered open scholarship more broadly. Melroy Almeida and Sarah Thomas highlighted the importance of persistent identifiers, such as ORCID IDs and RAiDs. Paul Arthur discussed a project studying digital representations of slavery, and Michael Falk presented on the affordances and limitations of Wikilambda, a Wikimedia project that seeks to build the technical architecture for facilitating translations of Wikipedia articles into any language. As part of work conducted by the INKE Partnership, Alan Colin-Arce presented on the importance of community governance for the Humanities and Social Sciences (HSS) Commons.
The third and final presentation session featured five talks. Nick Thieberger discussed PARADISEC and how his team has developed machine translation and speech-to-text tools with records of languages spoken in the Pacific region. Hamish Maxwell-Stewart described a project to visualize the lives of 75,000 convict lives in Tasmania based on historical sources and the assistance of AI image generation and avatars. James Smithies provided an update on the future plans for the HASS Digital Research Lab at the Australian National University.
Members of the INKE Partnership also presented in this session. Lynne Siemens discussed the ongoing collaborations at the INKE Partnership, highlighting that having the Electronic Textual Cultures Lab at the University of Victoria handle the administrative component of the partnership freed up time for other partners to focus on collaboration and research. In addition, Alan Colin-Arce presented on how the translations of the HSS Commons have helped build connections among researchers working in multiple languages.
Day 2
The second day of the CAPOS gathering kicked off with a plenary panel on how four different digital projects in Australia are addressing or considering the challenges posed by AI. Maggie Nolan from Auslit focused on bibliography, Tully Barnett from the Australian Creative Histories and Futures discussed cultural data, Samantha Bennett from the Australasian Council of Deans of Arts, Social Sciences and Humanities talked on higher degree research, and Nichola Burton from the HASS and Indigenous Research Data Commons discussed the provision of digital research infrastructure in Australia.
To close off the conference, Michael Sinatra gave a keynote where he proposed several research areas in which open social scholarship and AI can intersect, in particular by focusing on the mapping, accountability, and perspectives in AI. The talk concluded by posing the question of What would it mean to do open social scholarship with, and about, artificial intelligence.
Following the keynote, CAPOS partners met for a business meeting to discuss next steps to maintain the collaborations between Canadian and Australian partners and continue the conversations of the conference in 2026. Stay tuned to the INKE website for updates.
Recorded Presentations
In addition to the in-person presentations in Canberra, a few presentations were recorded for participants to view asynchronously, including two discussing work by members of the INKE Partnership and the Electronic Textual Cultures Lab (ETCL). These videos are available at the HSS Commons for anyone to watch. Feel free to leave a comment using the Questions or Reviews features.
The first presentation by Britt Amell titled Engaging with Platforms and Open Scholarship provides an overview of the Platforms research scan developed by members of the ETCL and reflects on the changes that have taken place since the scan was published at the beginning of 2025, including the risk of an increasing platformization of the scholarly communication ecosystem.
The second presentation by Faraz Forghan Parast titled Taking Bearings: Open Social Scholarship meets AIdescribes the current work by the ETCL team on a research scan on AI and open social scholarship.
The third presentation by Janet Catterall from Open Access Australasia discussed the implications of AI as it spreads across the open ecosystem. In the presentation, Catterall argues that openness demands technological transparency to assess and critique AI tools and models, as well as the development of AI technologies in collaboration with communities.
Responses from the CAPOS Partnership
Michael Falk (Senior Lecturer in Digital Studies, University of Melbourne):
The rise of Large Language Models (LLMs) has forced us to reconsider the value of openness. On the one hand, the old ideals of open scholarship remain as valuable as ever. Our research should be open to critique, and open to discussion. It remains essential to communicate our findings with a broad public, and to expose our data and analysis to fellow researchers. On the other hand, the extractive practices of AI companies make openness more costly. Does openness mean sharing all our research with model providers, so they can improve their products? The size of AI models also undermines the value of openness. What is the point of being ‘open’ with an AI model, if only someone with a supercomputer can run it? These questions can seem hard to answer, but the CAPOS meeting demonstrated that answers are possible. In the two keynotes especially, we saw how artists can repurpose the latest technologies for the purpose of inquiry and critique, and we saw how teams of humanistic scholars can work together to open up AI research to new kinds of discussion and debate. For all its problems, ‘openness’ remains a useful ideal in scholarly practice in the Age of AI.