Indeed. Though, there are are a few considerations ... you might find the resources at the Sentient Syllabus Project http://sentientsyllabus.org useful.
Have a look at our resources for the Sentient Syllabus Project, at http://sentientsyllabus.org - may be helpful.
Reacting to news posted today by @abcxyz @donwatkins @techhelpkb @johnew
---
While we were busy thinking things – it appears that the NYC Department of Education is embarking on a fascinating experiment with its staff and students https://www.zdnet.com/article/chatgpt-was-just-blocked-by-the-biggest-us-school-district-heres-why/
We will learn a lot from this.
For educators who are more convinced by the alternative, we have resources at the Sentient Syllabus Project http://sentientsyllabus.org
#SentientSyllabus #ChatGPT #AI #Education #HigherEd #University #Academia #syllabus
My take on banning #LLMs in scientific writing: Being a native speaker of English in academia is an unearned privilege that large language models have a chance to reduce. Labeling their use as plagiarism (as ICML does) is IMO insisting on this privilege. 🤷♂️ As a grad student, I struggled more with writing in proper English than with the actual content of my work. Using #ChatGPT to improve the language of my own ideas is no different than using a spell checker - it's not plagiarism.
LOL. Duplicated d, r, t, i, and (technically) j. But I love the creative approach.
Feeling your frustration. I actually like to quote John Warner "ChatGPT can't kill anything worth preserving". Because, I think too few realize how little will be left, including, it seems, Warner himself. The challenge is on to define how to build win win solutions that **evolve** from what we have. We founded the Sentient Syllabus project to craft resources: http://sentientsyllabus.org - and analysis on Substack.
We call this "Schrödinger Facts" in our recent substack post: https://sentientsyllabus.substack.com/p/chatgpts-achilles-heel
It's a consequence of the probabilistic way the algorithm works: facticity is not a part of the model. That will change ...
Exactly! We discuss the pitfalls, reasons ... and why this is still useful here: https://sentientsyllabus.substack.com/p/chatgpts-achilles-heel
There is, I think, indeed a philosophical question lurking there. ChatGPTs "understanding" is **implicit**, derived from the understanding of its training data. I think this kind of superposition is not something we have considered deeply before. Hope to get around writing a bit on that at the #SentientSyllabus project ... https://sentientsyllabus.substack.com
LOL ! However, concerning that it might pass the course, eminem style and all. We have some syllabus resources and analysis you might find useful at http://sentientsyllabus.org ... good luck with your term!
Good read - thanks Mark. I just posted a resources update on the Sentient Syllabus Substack - the Course Activities file has a section on "Critique and Improvement" that you might find relevant.
https://sentientsyllabus.substack.com/p/resource-updates-2023-01-05
#SentientSyllabus #ChatGPT #AI #Education #HigherEd #University #Academia #syllabus
You might be interested in what we are putting together at the Sentient Syllabus Project http://sentientsyllabus.org
and the analysis on Substack. Maybe going a bit deeper ... :-)
#SentientSyllabus #ChatGPT #AI #Education #HigherEd #University #Academia #syllabus
Exactly! We are discussing this in much detail at the Sentient Syllabus project – http::/sentientsyllabus.org has resources and analysis is on Substack.
#SentientSyllabus #ChatGPT #AI #Education #HigherEd #University #Academia #syllabus
ChatGPT has no ideas of its own, and is not a valid source. Follow the idea instead, track the sources. This may be a significant amount of work, we write about why this is useful here: https://sentientsyllabus.substack.com/p/chatgpts-achilles-heel
But sometimes you just want to cite the algorithm for its own sake. Cite that as a "synthesized communication" ... and follow the "personal communication" format of your favourite citation style.
HTH
#SentientSyllabus #ChatGPT #AI #Education #HigherEd #University #Academia #syllabus
Just finished a number of updates of our resource files at the Sentient Syllabus Project – summaries posted at https://sentientsyllabus.substack.com/p/resource-updates-2023-01-05. In brief: New sections on "Critique and Improvement" and "Personalized Tutoring" in the Course Activities file; most writing went into "Understanding AI Issues": training data and process, the Sysiphus work of plagiarism detection (hint: the tools are unlikely to work); first perspectives on energy use, machine use for better training - and the resulting "Kurzweilian" acceleration.
#SentientSyllabus #ChatGPT #AI #Education #HigherEd #University #Academia #syllabus
@Bookish@universeodon.com
I think you are right. At the http://sentientsyllabus.org Project we are approaching this from the perspective of HE. Fundamentally, we argue, we either learn to surpass the AI, or the AI will become a competitor, in a competition which we will lose. But how do we surpass it, when it is constantly improving? Part of the answer must be to use it as a stepping stone. But the other part is to "educate" society about the persistent value that we humans bring to the outcome, as a possessor, not just an emulator, of sentience.
Writers will play a crucial role in this discussion. Power to you!
Indeed - John Warner wrote: "#ChatGPT won't kill anything preserving".
So we applaud our new motivation to return to human interaction and pedagogy. But, does anyone have an idea how that will scale? And what will happen with the marginalized when we realize it doesn't scale?
That needs answers, and soon. cf. https://sentientsyllabus.substack.com
Welcome to have a look at the resources and analysis we have been writing up the Sentient Syllabus Project: here http://sentientsyllabus.org and on substack: https://sentientsyllabus.substack.com - the focus is specifically HE.
I discuss this in my latest post at https://sentientsyllabus.substack.com/p/chatgpts-achilles-heel Bottom line: it can actually be particularly useful for students if the AI is wrong in that way.