We love the Halloween spirit of OpenOregon with their webinar scheduled for October 31, a panel discussion on… get ready… Artificial Intelligence. It’s not scary, right?
Does AI give you the creeps? Join our panelists for a 90-minute conversation this Halloween to demystify current and future uses of AI in open education. We’ll explore current use-cases, copyright considerations, and do a live demo. Underneath the scary costume, AI is just another tool in the educator’s belt.
About our panelists:
- Rachel Bridgewater, Faculty Librarian, Portland Community College and co-facilitator of Copyright First Responders Pacific Northwest. Rachel is a public services librarian with a passion for teaching and sideline in all things copyright.
- Kim Ernstmeyer, Open RN Project Director, Chippewa Valley Technical College. Kim has been a nursing educator for almost 20 years and leads the Open RN project in developing OER textbooks and virtual simulations to promote student success and quality patient care.
- Dominic Slauson, Instructional Technologist, OpenRN; and Learning Experience Designer, University of California Irvine. Dominic is an instructional technologist and designer with over a decade of experience bringing innovative and engaging learning experiences to students.
- David Wiley, Chief Academic Officer at Lumen Learning, adjunct faculty in Instructional Psychology & Technology at Brigham Young University, and Entrepreneur in Residence at Marshall University.
Watch: Register for webinar
Yes, it would have been easy to toss a prompt into an AI generator for a scary image, but that’s so typical. How about an old fashioned, openly licensed, attributed photo?
BOO! flickr photo by cogdogblog shared into the public domain using Creative Commons Public Domain Dedication (CC0)
Thanks @hofera for hosting a fabulous webinar yesterday. I decided to NOT use AI to create my own image:
I’d keep my eye out on the OpenOregon site for the published recording and a heap of long links from an overactive chat. The presentation used is at https://tinyurl.com/2cwzet4j
- Very impressive to see use of AI generated images, voice, media for the case studies in the OpenRN simulations (done inside H5P branching scenarios, well worth exploring)
- Watch the replay for Dominic Slauson’s explanation of how these were done, especially when he did a live demo of using ChatGPT to generate a case study. See his site Codaptive Lab for great resources on AI prompts you can try directly as well as video examples of the AI generated videos used in the simulations, e.g.
This is hyperealistic, at the same time, like many, there is some kind of uncanny valley effect at work here. My own thought is that this is okay, it reinforces that this is a simulation. But watch to see the video above and consider how quickly this potential has evolved.
- Check the slides above as well for the ones Rachel Bridgewater of Copyright First Responders Pacific Northwest for “Some thoughts about copyright, AI, and OER” Rachel was not available, but the slides have key links, questions, and a lot of ¯_(ツ)_/¯ representing how much is not firm now. I like how she demonstrates the practice of attribution and how it is a multilayered approach.
- David Wiley provided a broad perspective on “OER and Generative AI: What can the future look like?” noting the potential in his opening framing thought, “The internet largely eliminated time and distance as barriers to education. Generative AI will largely eliminate access to expertise as a barrier to education.” suggesting that gAI may have an impact noted in the research on the Doer effect.
I liked that this session had a good chunk of time at the end for open and impromptu conversation. One small thing I noticed is that people shared that the AI search results from Bing provide their sources-- that was the easy target of problems with ChatGPT that many many of us noticed how it fabricated things that look like citations that were just approximations, and often not real. And its true, I just tried Bing search, and it was impressive that it provides links for references is returns to you. That i useful.
But that’s not the same as getting to the source of what the AI was trained on-- that, I understand is not really possible to traceback from a result. Maybe I am wrong (?) but I think there is a difference between getting accurate references in result (which is very good) and being able to source exactly where AI got to that result.
Thanks again OpenOregon for a relly worthwhile session.