Author: kaylea

We advocate for consideration of harms to communities (including online communities) as they respond to, are used for, and incorporate generative AI algorithms. One area of risk is one we call ‘Social Model Collapse,’ the notion that changes to community dynamics — social and technical — can disrupt the fundamental processes they rely on to sustain themselves and flourish. We see this as a clear point of shared concern among STS, HCI, Sociology, and Communication scholars, and are hosting an open panel at the upcoming meeting of the Society for Social Studies of Science (4S), September 3-7 in Seattle. This panel is being coordinated by CDSC members Kaylea Champion (University of Washington) and Sohyeon Hwang (Princeton University).

From our call for submissions:

Model collapse in machine learning refers to the deterioration such a model faces if it is re-fed with its own output, removing variation and generating poor output; in this panel, we extend this notion to ask in what ways the use of algorithmic output in place of human participation in social systems places those social systems at risk. Recent research findings in the generation of synthetic text using large language models have fed and been fed by a rush to extract value from, and engage with, online communities. Such communities include the discussion forum Reddit, the software development communities producing open source, the participants in the question and answer forum StackExchange, and the contributors to the online knowledge base Wikipedia.

The success of these communities depends on a range of social phenomena threatened by adoption of synthetic text generation as a modality replacing human authors. Newcomers who ask naive questions are a source of members and leaders, but may shift their inquiries to LLMs and never join the community as contributors. Software communities are to some extent reliant on a sense of generalized reciprocity to turn users into contributors; such appreciation may falter if their apparent benefactor is a tireless bot. Knowledge communities are dependent on human curation, inquiry, and effort to create new knowledge, which may be systemically diluted by the presence of purported participants who are only algorithms echoing back reconstructions of the others. Meanwhile, extractive technology firms profit from anyone still engaging in a genuine manner or following their own inquiries.

In this panel, we invite consideration of current forms of social model collapse driven by a rush of scientific-industrial activity, as well as reflection on past examples of social model collapse to better contextualize and understand our present moment.

Submissions are 250-word abstracts due January 31st; our panel is #223, “Risks of ‘Social Model Collapse’ in the Face of Scientific and Technological Advances” [Submission site link].

Kaylea Champion will be presenting a practitioner-focused talk to the Puget Sound Python user group (PuPPY) on December 18th at 5:30 p.m. PST at the GitHub offices in Bellevue, Washington. Details on how to attend this free event are available here. This practitioner-oriented talk will draw from her own research as well findings from across the academy.

If you are attending the ACM conference on Computer-supported Cooperative Work and Social Computing (CSCW) this year CSCW in San José, Costa Rica. You are warmly invited to join CDSC members during our talks and other scheduled events. Please come say hi!

This CDSC has four papers at CSCW, which we will be presenting over the next three days:

Monday: At 11:00 am in Talamanca, Kaylea Champion will be presenting “Life Histories of Taboo Knowledge Artifacts” (full paper)

Tuesday: At 11:00 am in Central 3, Zarine Kharazian will be presenting “Governance Capture in a Self-Governing Community: A Qualitative Comparison of the Croatian, Serbian, Bosnian, and Serbo-Croatian Wikipedias” (full paper, blog post), followed by Sohyeon Hwang presenting “Adopting third-party bots for managing online communities” (full paper, blog post)

Wednesday: At 2:30 pm in Guanacaste 3, Kaylea Champion will be co-presenting “Challenges in Restructuring Community-based Moderation” (full paper, preprint)

If you’re at CSCW, feel free to get in touch in person or via Discord!

In the seventh talk of the Science of Community track we organized for FOSSY, Google FOSS researcher Sophia Vargas offered an overview of different strategies for measuring the value of open source (particularly in the context of a company thinking about how to engage with FOSS).

Some of Sophia’s key insights are: models for measuring one-time cost are relatively widespread (but depend on outcome metrics like lines of code rather than difficulty); understanding the cost of maintenance and community is still in formative stages; and that business leaders can make use of research-grounded models developed to measure value and risk in an academic context into decisionmaking tools within a business context.

This is part 7 of an 8-part series sharing highlights from the Science of Community track at FOSSY. Visit the FOSSY site for more bio details and an abstract of the talk.

In the sixth talk of the Science of Community track we organized for FOSSY, independent FOSS researcher Darius Kazemi described the results of an interview study to learn from the moderation teams of decentralized social network servers. One of Darius’ key observations is the extensive compliance and legally-required work that running such a server requires.

This is part 6 of an 8-part series sharing highlights from the Science of Community track at FOSSY. Visit the FOSSY site for more bio details and an abstract of the talk.

In the final talk of the Science of Community track we organized for FOSSY, Computer Science professor and FOSS researcher Dr. Bogdan Vasilescu described his team’s work to understand how developers think about abandoned dependencies. One of the key insights from this work is that abandonment of dependencies is quite common, but that updating a package to remove the abandoned dependencies is very slow — and that one of the factors that drives faster replacement is when projects explicitly announce that they are ending maintenance.

This is part 8 of an 8-part series sharing highlights from the Science of Community track at FOSSY. Visit the FOSSY site for more bio details and an abstract of the talk.

This is part 5 of an 8-part series sharing highlights from the Science of Community track at FOSSY. Visit the FOSSY site for more bio details and an abstract of the talk.

In the fourth talk of the Science of Community track we organized for FOSSY, principal developer advocate Paige Cruz shared the results of her investigation into the subject of how we can all do a better job of asking questions of one another in FOSS communities. One of her key insights is to invite us to engage with the perspective of those who might answer our question, and to think critically about what details we include and whether they really help others understand and respond — for example, a screenshot of our code can’t be copy pasted and might be unreadable, but a screenshot of a UI bug might replace wordy description.

This is part 4 of an 8-part series sharing highlights from the Science of Community track at FOSSY. Visit the FOSSY site for more bio details and an abstract of the talk.

In the third talk of the Science of Community track we organized for FOSSY, FOSS leader Ben Ford described his experience navigating the changes in his role when the Puppet project’s commercial partner was acquired by a private equity company. One of the essential takeaways from this talk is the different perspective towards community that a FOSS company takes versus a private equity company, and the challenge of communicating value in this context.

This is part 3 of an 8-part series sharing highlights from the Science of Community track at FOSSY. Visit the FOSSY site for more bio details and an abstract of the talk.

In the second talk of the Science of Community track we organized for FOSSY, CDSC PhD student Matthew Gaughan shared his research to understand how communities actually use README and CONTRIBUTING documents. Although guides to FOSS communities often recommend these documents be extensive and used as part of welcoming new contributors, we find that READMEs are often quite preliminary, and that CONTRIBUTING guides are often a reaction to an influx of contributions.

This is part 2 of an 8-part series sharing highlights from the Science of Community track at FOSSY. Visit the FOSSY site for more bio details and an abstract of the talk.