{ "jason": "clark" }
banner
jasonclark.bsky.social
{ "jason": "clark" }
@jasonclark.bsky.social
librarian @msulibrary, professor, hacker, author, footballer, skateboarder, fly-fisher, dogwalker, coffee brewer - still trying to make fetch happen. www.jasonclark.info
December 18, 2024 at 6:24 AM
File under: Old, but new to me...

Tiny Sci-Fi stories from Pete Majarich
www.instagram.com/tinyscifi/?h...
petemajari.ch/About
December 13, 2024 at 3:14 AM
Humor and satire with too much truth. At times, it's hard for me to read The Onion and McSweeney's. Occasionally, I do find motivation. "We made algorithms that made your society increasingly impenetrable..." is actually a call to action for digital literacy.

www.mcsweeneys.net/articles/sor...
March 3, 2024 at 11:15 PM
@yscott.bsky.social presenting our research at AI4LAM on Responsible AI for Libraries.

Project and grant description:
www.lib.montana.edu/responsible-...

Slides:
docs.google.com/presentation...
November 17, 2023 at 8:09 PM
The Algorithmic Awareness research centered on algorithmic literacy by studying algorithms & learning enough to offer instruction, learning modules.

It also dipped into seamful design. A search engine prototype offering explanations of algorithms in action.

www.jasonclark.info/files/algori...
November 17, 2023 at 5:54 PM
Our work towards AI Literacy involves education about & making things with AI. Mike Ridley named "seamful design" - designing & making apparent when a system can be questioned or contested.

I'm continuing to think on ideas from earlier "Algorithmic Awareness" research. github.com/jasonclark/a...
November 17, 2023 at 5:41 PM
Starting AI4LAM today w/Mike Ridley speaking to Explainable AI priorities for LAM.

Human-Centered Perspectives in Explainable AI (HCXAI)
hcxai.jimdosite.com

Policy and Regulation

Algorithmic Literacy

Critical Making

No ranking in this list. I've continue to think about the literacy moment.
November 17, 2023 at 5:28 PM
Witness one way to approach transparency in this transfer learning demo I created. github.com/jasonclark/i...

Note how the model is giving a percentage to suggest confidence in the predicted metadata. It is clearly failing in this prediction. This is helpful/essential for users of the technology.
November 16, 2023 at 10:22 PM