Automoderator As An Example Of Community Driven Product Design Claudia Lo, Sam Walton Wikimedia Foundation, United States of America
Rushes to adopt the latest technologies to the field of community moderation are deeply lopsided, with volunteer community moderators left at serious disadvantages compared to deep-pocketed companies with access to expertise and capital. The closed-door nature of product development at the majority of tech companies means that the logic underlying the creation of new features is opaque. Volunteer moderators are recognized by said companies as a distinct stakeholder group, but one that is not necessarily given the ability to direct the future of the spaces in which they work. Meanwhile, orthodox product development processes are not well suited to meaningfully incorporating outside feedback at key stages of product development. What does this mean for those who want to equitably employ newer technologies in service of volunteer moderators? We present the ongoing development and deployment of the Wikimedia Foundation’s Automoderator product as an alternative vision. This paper will focus on the collaborative process being undertaken between the Moderator Tools product team at the Wikimedia Foundation and volunteer moderator communities to design and build Automoderator. It will cover not only the challenges of developing AI-powered tools in this manner, but also elucidate why such an approach remains rare in the industry. This is also a rare opportunity to lay out the process of product development, a process that is normally totally obscured from public view. Finally, we present Automoderator’s development process as a concrete example of collaborative product development, providing a model for future efforts.
Presentation 2
A Systematic Review of VirtualHumans.org and its Role in Virtual Influencer Research, 2019 to Present Jul Jeonghyun Park/Parke University of Toronto, Canada
In this paper, I conduct a systematic review of the use of VirtualHumans.org in academic studies of the virtual influencer industry. To do this, I analyze all references to and use of content from the Virtual Humans website in relevant published research from 2019 to 2024 (n=189), including the website’s database of existing virtual influencers, interviews with virtual influencer creators and virtual influencer characters themselves, articles on the state of the virtual influencer industry, and profiles of each listed virtual influencer. I pair this analysis with a brief history of VirtualHumans.org from a political economy perspective, noting the factors which went into the website’s creation; its acquisition by Offbeat Media Group, a digital marketing agency; and its organizational shifts following the sudden departure of its founder in 2023. Ultimately, I question whether academic viewpoints of the emergent virtual influencer industry, many which refer to the Virtual Humans website as a valuable resource for grasping a sense of the size and scope of the virtual influencer phenomenon, adequately consider the rooted biases and commercial interests represented by the website as not only a database but a powerful broker within the industry. Moreover, by narrating the organizational development of VirtualHumans.org as an enterprise, I contribute detailed context into the formation of these biases and commercial interests which inform its position in the virtual influencer scene.
