The PEAT team recently joined thousands of technology industry executives, entrepreneurs, end users, government officials, educators, and others at the 2019 CSUN Assistive Technology Conference. As always, this yearly event offered keen explorations into technology’s potential to break down barriers for people with disabilities in workplaces, community life, and other settings. This year’s CSUN sessions highlighted the sharp rise of artificial intelligence (AI) into everyday life, charting both the challenges and great potential that AI holds from an accessibility perspective.

Promise and Potential

Intuit’s Poonam Tathavadkar gave a fascinating discussion of the promise that AI can hold in facilitating more accessible content for people with disabilities. Software can now learn how to recognize and respond to images, sounds, and linguistic expressions, which has opened up new opportunities for people with a wide range of disabilities. Tathavadkar highlighted emerging trends for this space, such as:

  • Auto captioning with AI that can best the world’s top lip-reading experts by a ratio of 4:1
  • Autonomous cars—when built with Universal Design principles—can provide expanded transportation options for people who are currently unable to drive
  • Facial recognition and image recognition to support navigation and interaction with the environment for people who are blind or have low vision
  • Text summarization to enhance comprehension for people with cognitive disabilities

Microsoft’s AI for Accessibility initiative

From an enterprise perspective, Microsoft gave an exciting overview of their AI for Accessibility grants to accelerate the development of accessible and intelligent AI solutions. Microsoft’s top priorities for selecting grantees included technologies to address the unemployment rate for people with disabilities, which is currently double the rate for people without disabilities. Presenters Guy Barker, Wendy Chisholm, and Heather Dowdy noted that AI solutions could support job seekers and employees in developing their professional skills, shape improved workplace culture, and expand inclusive hiring.

The Challenges of Building Inclusive AI

Many CSUN sessions also highlighted key challenges and potential risks of AI. A particularly thought-provoking session featured an in-depth conversation among several AI researchers on the responsibilities researchers must prioritize related to privacy, ethics, and bias. Panel participants included Meredith Ringel Morris (Microsoft), Megan Lawrence (Microsoft) Matt Huenerfauth (Rochester Institute of Technology), Shiri Azenkot (Cornell Tech), and Jeffrey Bigham (Carnegie Mellon). Their remarks and commentary during the session emphasized drawbacks in the way AI uses training datasets to learn patterns and enhance predictions. For example:

  • Models learning from biased training data may reproduce historical bias against minority groups, including people with disabilities.
  • Training data typically underrepresents outlier populations, including people with disabilities, and frequently does not perform as well for these groups.
  • Gathering inclusive datasets will prove essential for building effective solutions, but also holds significant ethical challenges. Building these systems requires people to waive privacy rights, and people with disabilities may have heightened privacy concerns.

In another CSUN session, Phill Jenkins and Erich Manser shared an intriguing overview of IBM’s activities in AI bias detection and mitigation techniques. IBM’s AI Fairness 360 toolbox includes resources for testing data, models, and outcomes for bias. Like other presenters, they similarly identified the challenge of building inclusive datasets as one of the largest looming fairness challenges for researchers.

As always, PEAT was delighted to be part of the CSUN proceedings, including presenting our own sessions highlighting our Staff Training Resources and our work with Teach Access to close the Accessible Technology Skills Gap. Both of these tools arose from needs identified at CSUN in years past, and the emerging trends we identify at the conference each year continue to guide our work. Please be sure to check out the 2019 season of our Future of Work podcast series to stay on top of the latest issues in emerging technology, accessibility, and workplace inclusion!