Hanlin Li

Online communities and their host platforms are mutually dependent yet conflict-prone. When platform policies clash with community values, communities have resisted through strikes, blackouts, and even migration to other platforms. Through such collective actions, communities have sometimes won concessions, but these have frequently proved to be temporary. Although previous research has investigated strike events and migration chains, the processes by which community-platform conflict unfolds remain obscure. How do community-platform relationships deteriorate? How do communities organize collective action? How do the participants proceed in the aftermath? We investigate a conflict between the Stack Exchange platform and community that occurred in 2023 around an emergency arising from the release of large language models (LLMs). Based on a qualitative thematic analysis of 2,070 messages from Meta Stack Exchange and 14 interviews with community members, we reveal how the 2023 conflict was preceded by a long-term deterioration in the community-platform relationship, driven in particular by the platform’s disregard for the community’s highly valued participatory role in governance. Moreover, the platform’s policy response to LLMs aggravated the community’s sense of crisis, triggering strike mobilization. We analyze how the mobilization was coordinated through a tiered leadership and communication structure, as well as how community members pivoted in the aftermath. Building on recent theoretical scholarship in social computing, we use Hirschman’s exit, voice, and loyalty framework to theorize the challenges of community-platform relations evinced in our data. Finally, we recommend ways that platforms and communities can institute participatory governance to be durable and effective.

Social media users create immense value for various technologies such as search engines, content platforms, and most recently, generative AI models. Among social media users, those who produce content professionally, such as artists and creative workers, have expressed strong concerns about unauthorized reuse of their content; however, it remains unclear how the vast majority of casual users perceive content reuse, in particular for AI training. To gain a comprehensive understanding of users' perspectives, we surveyed 478 Instagram users from a university network about how much they value their content for commercial reuse and their acceptance of specific content reuse scenarios. We found that under reverse auction, in which the lowest bidder wins, respondents' evaluation of their content for commercial reuse is highly skewed, with a median of $62 and a long tail of more than thousands of dollars. A nontrivial percentage of respondents found content reuse for search engine display, education, and academic research acceptable. Conversely, content reuse for commercial or non-profit generative AI models was a negotiable scenario for most respondents. Most notably, compensation was likely to make almost all scenarios more acceptable to users. Our work highlights the diverse, contextually dependent perspectives that users have. We discuss how design, policy, and market practices might support contextually aware content reuse.

Wikidata, an open structured database and a sibling project to Wikipedia, has recently become an important platform for information professionals to share structured metadata from their memory institutions-organizations that maintain public knowledge and cultural heritage materials. While studies have investigated why and how peer producers contribute to Wikidata, the institutional motivations and practices of these organizations are less understood. Given Wikidata's potential role in linking and supporting knowledge infrastructures and open data systems, we examined why and how information professionals in memory institutions use Wikidata as part of their organizational workflow. Through interviews with 15 participants, we identified the three archetypal roles of Wikidata users within memory institutions-providers, acquirers, and mutualists-and the different types of contributions that these institutions bring to Wikidata. We then explored potential collaboration opportunities between memory institutions and other volunteers in Wikidata, discussed the value of the data work conducted by these professionals, and examined how and why they track their contributions. Our work contributes to the wider discussions around collaboration and data work in CSCW by 1) studying the motivations and practices of information professionals, their differences from those doing volunteer work, and opportunities for the Wikidata community to promote more collaborative efforts within memory institutions and with other volunteers and 2) drawing attention to the important data work done by memory institutions on Wikidata and pointing out opportunities to support the contributions of information professionals.

This SIG will engage CSCW researchers in a collaborative discussion on the use of UGC from social platforms for academic research. In recent years, both the importance and difficulty of accessing such data for scientific inquiry have increased. Drawing on early experiences from Reddit's 'Reddit for Researchers' data-sharing initiative, this session will: (1) identify key directions in social computing research that such initiatives can support, (2) enumerate ethical and user safety concerns which must be addressed, and (3) ideate governance models to enable community-driven review of research data requests. Participants will inform the design of a more inclusive and effective research environment for the study of social platforms, better bridging the gap between industry and academia.

Content creators' work has long been reused by various technology developers and researchers; recently joining the ranks of content reusers are emergent generative AI models. However, creators' perspectives and acceptances of these use cases were often overlooked in the process. We present our initial findings on how content creators perceive scenarios of content reuse as acceptable, negotiable, or unacceptable. We surveyed 478 content creators on Instagram and a non-trivial percentage of respondents found content reuse for search engine display, education, and academic research acceptable. Conversely, content reuse for commercial or non-profit generative AI models was a negotiable scenario for most respondents. Receiving compensation was likely to make almost all use cases more acceptable to content creators. Our work paves the way for future work into understanding how creators value their work, research ethics, and best practices for using user-generated content for responsible technological innovation.

Large language models (LLMs) stand to reshape traditional methods of working with data. While LLMs unlock new and potentially useful ways of interfacing with data, their use in research processes requires methodological and critical evaluation. In this workshop, we seek to gather a community of HCI researchers interested in navigating the responsible integration of LLMs into data work: data collection, processing, and analysis. We aim to create an understanding of how LLMs are being used to work with data in HCI research, and document the early challenges and concerns that have arisen. Together, we will outline a research agenda on using LLMs as research tools to work with data by defining the open empirical and ethical evaluation questions and thus contribute to setting norms in the community. We believe CHI to be the ideal place to address these questions due to the methodologically diverse researcher attendees, the prevalence of HCI research on human interaction with new computing and data paradigms, and the community’s sense of ethics and care. Insights from this forum can contribute to other research communities grappling with related questions.

The rise of prominent AI models such as ChatGPT and Stable Diffusion has brought the scale of commercial web scraping to the forefront attention of content creators and researchers. Billions of webpages and images are used to train these models without content creators’ knowledge, sparking extensive criticism and even lawsuits against AI firms. Amidst such debates, licensing is proposed by researchers and legal experts to be a potential approach to mitigate content creators’ concerns and promote more responsible data reuse. However, it remains unclear what specific licensing terms will be effective to mitigate content creators’ concerns and what sociotechnical environments are necessary to facilitate the use of licensing at scale. This workshop will provide a venue for researchers, content creators, and legal experts to answer these questions.

Many recent technological advances (e.g. ChatGPT and search engines) are possible only because of massive amounts of user-generated data produced through user interactions with computing systems or scraped from the web (e.g. behavior logs, user-generated content, and artwork). However, data producers have little say in what data is captured, how it is used, or who it benefits. Organizations with the ability to access and process this data, e.g. OpenAI and Google, possess immense power in shaping the technology landscape. By synthesizing related literature that reconceptualizes the production of data for computing as “data labor”, we outline opportunities for researchers, policymakers, and activists to empower data producers in their relationship with tech companies, e.g advocating for transparency about data reuse, creating feedback channels between data producers and companies, and potentially developing mechanisms to share data’s revenue more broadly. In doing so, we characterize data labor with six important dimensions - legibility, end-use awareness, collaboration requirement, openness, replaceability, and livelihood overlap - based on the parallels between data labor and various other types of labor in the computing literature.