Participatory AI Research Network
How can community organizers, journalists, nonprofits, and academics partner with the public to develop evidence to improve people’s digital lives? That’s the motivating question behind a new community of practice at the Coalition for Independent Technology Research.
Part of the Coalition’s purpose is to support community efforts at growing the movement of people who work to bring reliable evidence to questions about technology and society. So when over a hundred people from around the world reached out in response to a paper by Megan Price and J. Nathan Matias on public involvement in the science of AI, we knew there was an opportunity to realize the collective potential of so many researchers.
Thanks to a stellar, generous organizing team, we are now holding monthly gatherings to share ideas, get feedback, and offer mutual aid on projects. We think of this meet-up as complementary to other initiatives. If you’re interested in policy and governance, PAIRS is a great community. If you care about design and technology making, PDC provides a creative scholarly environment. Many of us routinely connect with the Association for Advancing Participatory Sciences, who provides a support community around big-picture questions of public participation in science. Our monthly meetings are designed to provide peer support with on the ground projects and methods ideas for projects focused on evidence-collection, journalism, science, and evaluation.
To get to know each other better, we spent our June meeting asking folks to share a project that inspires them – by someone other than themselves, whether or not it involved AI. Here are some of the examples our community shared.
Nine Participatory Projects that Inspire Us:
- Organized by Lily Irani, Six Silberman, and Amazon Mechanical Turk workers
- Shared by J. Nathan Matias
- Turkopticon organizes mutual aid, collects data, and carries out advocacy to make digital work on Amazon Mechanical Turk a good job. They started out with a browser plugin that made it easy for workers to collect shared information and ratings about job-requesters in order to make smart choices about which jobs to take. Over time, that data turned out to be useful for understanding Amazon’s platforms more directly.
- I admire the way Turkopticon organized “mutual aid accountability” where data collection met a very clear need for workers while also producing evidence that could contribute to change and also advance scientific understanding of the labor behind AI systems. I wrote about Turkopticon in an article for the Atlantic in 2015, and the example helped inspire me to start the Citizens and Technology Lab.
Worker’s Algorithmic Observatory
- Organized by Abani Ahmed, Christine Ao, Dana Calacci, Felix Chen, Clayton, Samantha Dalal, Ananya Gollakota, Amado Krsul, Varun Mani, Caroline McLaughlin, Andrés Monroy-Hernández, Ryan Oet, Varun Rao, Cindy Tong, Eesha Agarwal, Youjean Cho, Cathy Di, Mihir Kshirsagar, Amna Liaqat, Eduardo Moreno, Serhiy Ozhbiko, Kok-Wei Pua, Andrew Schwartz, Satya Shodhaka, Danny Spitzberg, and Tobey Switzer
- Shared by Alan Smith
- The WAO is a collection of academics at multiple institutions who are working together to solve the problem of how to best investigate black-box platform algorithms, such as those used by Uber and DoorDash. They develop tools and support for workers and allies to crowdsource data and investigate the black-box algorithmic systems behind the platforms that determine pay, schedule, and more, in the platform economy and beyond.
- I admire their work for many reasons, but I’ll focus on two: Their focus on giving voice to the labor side of these platforms, and their longitudinal look at driver-side data. Because they’ve just been sticking with it for so long, they’ve created an incredably valuable database of rideshare rides from the driver-side that a single snapshot can never do… and was very valuable in my work!
Participatory Harm Auditing Workbenches and Methodologies Project
- Organized by a team of researchers across the UK
- PHAWM is creating processes for participatory auditing and training stakeholders in how to carry out paticipatory audits of predictive systems and generative AI.
- Organized by George Aye and Sara Cantor
- The Greater Good Studio advances people-centered social change by working alongside people who are impacted by social inequity. They have worked on a wide range of issues including rural childcare, barriers to black and latino home ownership, and many other issues.
- They are fantastic at participatory design and collaborating with communities. They are also excellent at attending to power differentials in the design process.
- Organized by a federation of innovators
- Shared by Megan Price
- The federation is a network of smaller / individual technology companies that are leveraging local and hyper specific language knowledge to develop task-specific models that serve their needs. The federation is facilitated by DAIR.
- They are focusing on small, community AI rather than the one model that solves all things for all people. They are providing an example for how to think smaller in a way that is effective.
- Masakhane is a grassroots organisation whose mission is to strengthen and spur NLP research in African languages, for Africans, by Africans.
- Their approach is community-driven. They are trying to improve the massive gap that exists for big languages spoken by hundreds of millions of people, led by people who speak that language. They also are bridging between language community leadership and scholarly knowledge production in linguistics and computer science.
- MindBench is working to improve the evaluation of LLMs in mental health contexts, through partnerships across clinicians, researchers, technologists, and people with lived experience of mental illness. MindBench works with the National Alliance on Mental Illness and Beth Israel Deaconess Medical Center’s Division of Digital Psychiatry.
- MindBench provides easy ways for people with a history of mental illness to volunteer through their crowdsourced rating program, as well as more in-depth opportunities to shape the overall direction of the project.
Seasonal Medicine Wheel Project
- Organized by Sonya Cressman
- Shared by Brian Fisher
- The Seasonal Medicine Wheel project partnered public health researchers with Witsuwit’en Elders, Scholars and knowledge keepers who live in what is currently Western Canada to co-design risk communication materials, as part of a project on Artificial Intelligence in Technology Enabled Clinical Care.
- Organized initially by Tom Harrisson, Charles Madge, and Humphrey Jennings
- Recommended by Sarah Allen
- People were encouraged to keep diaries of their everyday lives to create an “anthropology of ourselves” in the UK. The project started in UK in 1937 and it was continued during WWII. The archive became a great resource on middle class life in the mid-century. Follow-ups took place in the 80s and during the COVID-19 pandemic.
- The project provided a unique view into middle class life in the period of great social change, with many women entering the workforce for the first time. People tracked everything from overheard pub conversations and shopping habits to how people decorated their homes.