Autumn 2024 - Spring 2025
Research into the Harms caused by GAI Tools
This year-long DRG will explore various facets of working with Generative AI tools, with a specific focus on the ways in which their outputs can and do cause harm towards traditionally marginalized populations. Led by doctoral candidate Sourojit Ghosh, students in this DRG will be expected to conduct impactful novel research in this field and submit high-quality work to relevant conferences for publication. The DRG will have biweekly large-group meetings on Tuesdays from 12-1:30 p.m. (Fall quarter, Winter and Spring times TBD), as well as small group meetings scheduled throughout the year. Students can enroll for a maximum of 5 credits throughout the academic year, distributing them as they see fit over three quarters.
Ideal students will have some experience working with Generative AI tools, with prior knowledge of computational and/or human-subjects evaluations of their outputs preferred. Students must commit to participating for the entire academic year, as well as being available for the initial meeting time listed above. We are looking for a maximum of 6 students. Interested applicants can fill out this Google Form, or reach out to ghosh100@uw.edu for questions.
Spring 2024
Exploring Engagement with Virtual Pet Sites
Organizers: Alyse Marie Allred, Cecilia Aragon
Note this DRG is full for Spring 2024 and no longer accepting applications.
Virtual pet sites, such as Neopets, are browser games built around the core mechanic of collecting digital pets, often with additional features such as: minigames, forums, contests, dress-ups, and a tradesmarket. Although they rose to prominence in the early 2000s, virtual pet sites persist to this day. Moreover, while the original trend was associated primarily with children, many of the current users of these sites are adults--some of whom were the original children who have since grown up. This DRG seeks to understand how adults interact and play with these virtual pet sites, as a point of comparison to existing ethnographic data on how children engage with them. The goal is to produce a paper summarizing the findings and submitting it to a conference or journal. Co-authorship on the paper is a possibility for motivated and engaged students.
What students should expect:
- Joining one of four virtual pet sites (Neopets, Flight Rising, Dappervolk, Lorwolf)
- Independently playing at least 30 minutes, three days of the week (1.5 hours playtime total)
- Weekly short reflections (1-3 paragraphs) relaying activities, observations, and other significant interactions
- Weekly 2 hour meetings with the team to share work and discuss observations (meeting time TBD)
- Full reflections at midterms and finals
For questions, please contact: Alyse Allred at alyse.allred@gmail.com.
Autumn 2023 - Spring 2024
Human-Centered Data Science and Large Language Models
This year-long Directed Research Group will explore questions in the field of human-centered data science, as it relates to the development, usability, evaluation and social impacts of the recent proliferation of large language models. Students will lead and participate in original research projects, as they discuss novel and impactful questions in the field by designing and executing their own studies. Students will be working directly with PhD candidate Sourojit Ghosh (G).
Students have the option of signing up for 4 credits, across three quarters. In the Fall 23 quarter, the DRG will meet Tuesdays 12-1pm, and students will be required to have an additional 1-on-1 meeting with their research team once a week. To apply, please fill out this form by EOD September 17. For questions, contact G at ghosh100@uw.edu.