[Athen] Harnessing Copilot and other gen-AI tools to increase digital accessibility efficiency

[Matthew Deeprose via athen-list]

[With apologies for cross-posting]

Could generative AI tools like Microsoft Copilot<https://copilot.microsoft.com/> reduce effort or workload with any Higher Education digital accessibility workflows?

I've put together a video presentation<https://matthewdeeprose.github.io/harnessCopilot.html> with a corresponding written transcript<https://matthewdeeprose.github.io/harnessCopilot__transcript.html> that cover four scenarios based on my experience of using Copilot and similar tools in my role for the following tasks over the past six months:

  *   Converting captions into a formatted transcript.
  *   Writing alternative text for images.
  *   Converting a podcast in mp3 format into a formatted transcript with speaker attribution.
  *   Creating sample content for use in accessibility remediation workshops.

In the presentation, I'll introduce each concept and explain why it's important. I will demonstrate Copilot in use and share the prompts I created.

I also provide some introductory information about Copilot for those who have not used it before.

Of course, there are concerns around bias and the impact of the climate of generative AI. I'll raise those in the presentation.

The presentation concludes with a set of questions for Higher Education institutions to consider as they start to use generative AI for digital accessibility workflows.

While these scenarios are fairly niche and limited, I hope that others will continue the conversation and share their experiences.

Happy to receive feedback and hope to learn how you have been using tools like this in the digital accessibility sphere.

Matthew Deeprose
University of Southampton

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman12.u.washington.edu/pipermail/athen-list/attachments/20240904/ac65bcc8/attachment.html>