Navigating Privacy Issues with Generative AI Panel

On March 7th, I was thrilled to participate in a Navigating Privacy Issues with Generative AI panel hosted by the Idaho State Board of Education and moderated by Jen Schneider from Boise State University.

The Panel and Curated Privacy Resources for Educators

The other folks invited to be on the panel were Liza Long from the College of Western Idaho and Reed Hepler from the College of Southern Idaho, who each have extensive experience using and teaching with AI tools.

To complement the panel, we curated a list of related Privacy Resources for Educators.

My Perspective

I described my approach as more of an “enthusiastic skeptic,” informed by my background with media studies and digital humanities tools. Automation certainly has great potential when its results are verifiable and accurate—but I’m far from convinced that the currently-popular tools live up to the hype about their utility, that their outputs are worth the material trade-offs involved in data storage, or that the companies involved have earned our trust.

I also think that we should promote a media literacy-informed approach to these tools, one that deeply considers how the claims generated by these tools include biases and that also questions who benefits from a “free to use” technology. (Spoiler: I reference Richard Serra and Carlota Fay Schoolman’s classic 1973 piece “Television Delivers People” to point out that, much as t.v. delivers viewers to advertisers, current AI tools deliver users and their associated data to capitalist surveillance companies.)

With a more optimistic view, though, I hope this “AI hype moment” provides us opportunities to highlight the various ways of knowing that people can learn through higher education. We also can use this to have more conversations about the human purposes and human values of what has traditionally been called “liberal education”: an education that aims to help people prepare themselves for an uncertain future, to consider what they think of as a good and proper life, and to engage deeply in the rights and responsibilities we have as citizens in free, democratic societies.

I’m also cautiously optimistic about using this moment to educate people about cloud computing generally, as well as the specifics of how these AI tools function through pattern recognition, generating remixes of sources, and producing plausible rather than accurate outputs. As I recently wrote, it’s a great opportunity for educators to examine their assumptions about languaging and knowing.

Similarly, I’m looking forward to variations on these tools that allow users to run them locally, to choose their own training data, and to educators figuring out ways to provide privacy-first access to these tools.

Current Accessibility and Potential Future Refinements

The video of our panel is closed-captioned and there is a transcript available on the YouTube site. I haven’t yet read it closely enough to spot any errors, but I’ll try to do so in the next couple of weeks.

I may also try to make more bullet-point versions of at least my own contributions, as I’m personally someone who always prefers reading a transcript (sometimes even a summary) over watching a video.

Kudos

Did you enjoy this? Let me know:

Perhaps even leave a comment below?