On Tools, AI, and Disability

What do we mean when we say that "AI is a tool" and what does disability have to do with it? In this talk, I will introduce my prior work as a tool-user, tool-maker, and researcher of tool-makers at the intersection of interactive data science and accessibility. I have focused my PhD on "tool-making," and specifically by using tool-making as an intervention on the accessibility of data experiences: by helping practitioners consider people with disabilities, creating low-level building blocks and scaffolding tools and infrastructures, and by questioning who "practitioners" are.

My future agenda now investigates how disability, AI, and "tools" are entangled. I am interested in 3 areas of exploration: First, remaining embedded with my communities and continuing to advance the state of the art for open source tooling in data science and accessibility. Next, I am interested in investigating ways that modern AI "piggybacks" on accessibility and ways that assistive technologies become a backdoor for model-building and data collection. Lastly, I want to interrogate ways that AI as a tool is commonly justified (economically and ethically) both by relying on the dynamics of a user-who-is-disabled and through "access washing."

Frank is a PhD candidate and researcher at the Human-Computer Interaction Institute at Carnegie Mellon University. His work explores the intersection of technical toolmaking, interactive data science, and accessibility. Frank has been visualizing data professionally for over a decade and his contributions are used internationally by over 15 policy and governance orgs (including the European Commission and World Health Organization), in over 20 undergraduate and graduate courses, and at over 110 companies, including 3 of the "Fortune 5." Frank has had a wide array of professional, industry collaborations with corporations like Apple, Adobe, Microsoft, and Visa, to open-source data science and standards communities like Bokeh, Quansight, Anaconda, Highsoft, Vega-lite, and the World Wide Web Consortium's Web Accessibility Initiative. His research projects include Chartability, a framework for helping practitioners evaluate inaccessible visualizations, Data Navigator, a tool that helps developers make interactive visualizations more accessible for users of assistive technologies, and Softerware, a system-design approach that helps guide architects and ecosystem engineers towards personalization and end-user agency in software systems.