Informing an AI Ethics of Accessibility

October 28, 2021
3:00 - 4:00 pm ET
Halligan 102, or Zoom
Host: Elaine Short

Abstract

Artificial intelligence (AI) promises to scale accessibility solutions like never before. For example, machine learning, computer vision, and natural language processing techniques may convert visual information including documents, photos, and charts into nonvisual formats, making them quickly accessible to screen reader users and people who are blind or who have low vision. However, the contents of this information is not neutral and research and media reports continue to illuminate AI- bias and malicious applications. These harms tend to impact people who already experience marginalization based on their race, gender, or disabilities, complicating the ethics of rapid deployment of AI solutions meant to increase accessibility.

In this talk I will overview two projects examining AI applied to address inaccessibility and inequity in disparate contexts. The first concerns representations of race, gender, and disability in descriptions of photographed people and the second concerns emerging networked and automated transportation technologies meant to increase mobility options for people and goods delivery. In both projects, I took up a justice perspective by not only centering people who might benefit from the promised access of these technologies but also those who are likely to experience harm based on known AI-bias. With my empirical studies, I will point to future directions for an AI ethics to guide emerging and automated accessibility solutions.

Bio:

Cynthia Bennett is a Postdoctoral Fellow at Carnegie Mellon University’s Human-Computer Interaction Institute and a Researcher in Apple’s AI and Machine Learning organization. Her HCI research concerns the intersection of power, disability, design, and accessibility. She positions the lived experiences and creativity of people with disabilities as starting points for designing accessible and justice-oriented technology. Bennett is regularly invited to speak about her research; recent hosts include The Good Robot podcast and Apple’s Worldwide Developers Conference. She has received funding from the National Science Foundation, Microsoft Research, and the University of Washington’s Human Centered Design and Engineering department, where she completed her Ph.D. She has published in top-tier human-computer interaction venues, and six of these papers have received awards.

Please join meeting in Halligan 102 or via Zoom.

Join Zoom Meeting: https://tufts.zoom.us/j/97183120811

Meeting ID: 971 8312 0811

Password: See colloquium email

Dial by your location: +1 646 558 8656 US (New York)

Meeting ID: 971 8312 0811

Passcode: See colloquium email