Boris Steipe @boris_steipe@qoto.org
- Sentient Syllabus Project
- http://sentientsyllabus.org
- Sentient Syllabus Analyses
- https://sentientsyllabus.substack.com
Joined Nov 2022
Boris Steipe
boosted
I compiled some #ChatGPT in #education resources, in case it is of use, especially for any #EdDev folks planning their own workshops on the topic: https://edtechdev.wordpress.com/2023/01/31/planning-a-workshop-on-ai-tools-like-chatgpt-for-your-school/
#AI #EdTech #FacDev @edutooters
Boris Steipe
boosted
Re #ChatGPT:
Some #professors seek to craft assignments that guide #students in surpassing what #AI can do.
Others see that as a fool’s errand—one that lends too much agency to the software.
My @insidehighered story. #highereducation #highered
https://www.insidehighered.com/news/2023/01/31/chatgpt-sparks-debate-how-design-student-assignments-now
Very insightful coverage by @susan_dagostino in #InsideHigherEd "Designing Assignments in the ChatGPT Era". If only more of our discourse could have that level...
Though: disclaimer: I had the honour to add a few ideas I wrote for the Sentient Syllabus project http://sentientsyllabus.org
Boris Steipe
boosted
An analysis of the automatic bug fixing performance of ChatGPT. ~ Dominik Sobania, Martin Briesch, Carol Hanna, Justyna Petke. https://arxiv.org/abs/2301.08653 #AI #ChatGPT #Programming
Boris Steipe
boosted
"The big take-home lesson was that all these studies were generally considered acceptable by the expert reviewers. This is rather astounding: a chatbot was deemed capable of generating quality academic research ideas. This raises fundamental questions around the meaning of creativity and ownership of creative ideas — questions to which nobody yet has solid answers."
Boris Steipe
boosted
Nun, #Chatgpt ist ja nicht auf den Kopf gefallen. Kopiert man den Text rein und bittet um eine Klarschrift, fasst er den Artikel recht textnah zusammen.
Boris Steipe
boosted
"Sprachmodelle wie #ChatGPT machen einmal mehr deutlich, dass stupides Auswendiglernen und die Abfrage von derartigem Wissen an Schulen und Unis ein überholtes Konzept sind", sagt Schmid.
Boris Steipe
boosted
Can #ChatGPT take the pain out of annual academic reviews? I thought I'd find out.
This is my very long and productive collaboration with ChatGPT on a process I really dislike!
First time I actually was laughing with ChatGPT. There's this announcement that today's upgrade of the system came with improved factuality. https://help.openai.com/en/articles/6825453-chatgpt-release-notes
Of course I had to test it. Nice. It can continue Fibonacci sequences, and it can continue 1, 0.5, 0.25, correctly. But it estimates rather than computes successive square-roots and can't really explain what it was doing there. Whatever.
Then I ask it for sources for the comparison between "specifications grading" and the "ungrading" movement. I get five books ... none of which actually exist with exactly those title, authors and years. My, my. Though this is not entirely useless, as always since: ...
Human: ... if only those existed. But yes, similar titles float around in that field, and some of those authors have made their careers with the topic. Thank you.
ChatGPT: You're welcome! I apologize for the confusion caused by the incorrect sources. It's great to hear that similar titles and authors do exist and are actively working on these topics.
----
Oh my! Indeed. That's a relief to both of us. 😂😂😂 Similar titles exist.
----
I'm sure factuality has improved if they say so. I'm also sure that there's scope for more improvement. Like an actual search for sources. Two more months, right?
Boris Steipe
boosted
A new ChatGPT was released today:
"We’ve upgraded the ChatGPT model with improved factuality and mathematical capabilities."
Source: https://help.openai.com/en/articles/6825453-chatgpt-release-notes
I think ChatGPT works great when it comes to generating believable text, but math is (still) not its strength.
Norms for Publishing Work Created with AI https://dailynous.com/2023/01/30/norms-for-publishing-work-created-with-ai/
The dailynous picks up on the topic we covered two days ago at https://sentientsyllabus.substack.com/p/silicone-coauthors ... with a discussion that is a bit deeper than many others. Still, much more to be said.
Boris Steipe
boosted
Boris Steipe
boosted
New Blog Post. "Education as We Know it is Over." http://edtechman.blogspot.com/2023/01/education-as-we-know-it-is-over.html
#edtech #edchat #edutooter #edutooters @donwatkins I did now use an AI to write this. @edutooters #education #chatgpt #artificialintelligence #machinelearning
Boris Steipe
boosted
Love this piece on the role of #emotion in #science & #research!! So many thoughts, but YES decision & action are inextricably tied to #emotions, including for scientists. When people talk about rational #logic they often assemble 'facts' to confirm their own emotion-based biases. Acknowledging the role of emotions in research improves our science. @BPSOfficial https://www.bps.org.uk/psychologist/emotions-essential-part-science
Boris Steipe
boosted
The dangers of loyalty in the workplace:
New research finds that, “instead of protecting or rewarding them, loyal employees are selectively and ironically targeted by managers for exploitative practices.”
🔏 https://doi.org/10.1016/j.jesp.2022.104442
Two quotes follow: 🧵👉
#Science
#Psychology
@psychology
#SocialPsychology
#SocialScience
@socialpsych
#OrgBehavior
@orgbehavior
#Loyalty
#Exploitation
#WorkplaceRelations
#Ethics
Getting #AIauthorship right is harder than one would think.
I just posted an analysis on AI authorship over at #SentientSyllabus:
https://sentientsyllabus.substack.com/p/silicone-coauthors
Key takeaways include: that arguments based on our usual criteria of /contribution/ and /accountability/ are brittle; that the problem lies with authorship being a vague term (cf. sorites paradox); that we are using posterior reasoning to justify our intuitions; and that reliable intuitions about the actual nature of the emergent(!) source-AI-author system need more work. A practical policy proposal rounds it off: empower the authors, use meaningful acknowledgements, quantify contributions.
Reuters:
At a time when Google and Bing are gearing up for the tech showdown of the decade ...
they are laying off five and six percent of their (human) workforce.
https://www.reuters.com/business/google-parent-lay-off-12000-workers-memo-2023-01-20
#SentientSyllabus #ChatGPT #AI #Education #HigherEd #University #Academia #Google #Alphabet #Microsoft #Bing
Boris Steipe
boosted
The recent debate in Norway about Chat GPT made me revive my old blog, and write a text about school as a tardigrade. We need to kill it before AI does. Maybe it is relevant for other countries as well, so I have translated it into English. https://margreta.wordpress.com/2023/01/17/is-school-a-tardigrade/
#AI #education #school #chatgpt
Boris Steipe
boosted
"More than 6,000 teachers from Harvard University, Yale University, the University of Rhode Island and others have also signed up to use GPTZero, a program that promises to quickly detect A.I.-generated text, said Edward Tian, its creator and a senior at Princeton University."
The wrong approach imo. One should focus on teaching how to work with #AI. Not on bans, the enforcement of which is unrealistic anyway. #ChatGPT
Boris Steipe
boosted
Clear and reasonable discussion of how to address #ChatGPT in teaching, by @andrewpiper (but posted elsewhere alas): https://twitter.com/_akpiper/status/1615034533232902151
- Sentient Syllabus Project
- http://sentientsyllabus.org
- Sentient Syllabus Analyses
- https://sentientsyllabus.substack.com
Joined Nov 2022
