aewolfai.bsky.social
I'll be posting my thoughts on various AI platforms for teachers and promoting good uses of LLMs. Read my thesis https://repository.usfca.edu/thes/1547/
56 posts
26 followers
16 following
Regular Contributor
Active Commenter
comment in response to
post
Or that he's using 20 yos to bully public officials?
comment in response to
post
I agree with this as well. We should connect. My thesis was in part written because of the hype around AI without tools being created.
There are some tools out there now with specific use cases, but we need more of them and we need better guardrails and context for them.
comment in response to
post
I agree. What needs to be in front of the education practice is the ability for expert teachers to evaluate how AI as a tool benefits or detriments the learning objectives and their pedagogy. Any other optimization that doesn't keep that in mind will lead to negative outcomes for the kids.
comment in response to
post
Being honest doesn't get you $500bn in investment unfortunately.
comment in response to
post
This might be an absurd question from the otherside of the academic sphere, but isn't half of the problem in terms of ML/LLM data sets the lack of data and the corruption of data? Why haven't there been data models that fix that? Why can't we automate the allegedly low-skill labor to generate data?
comment in response to
post
Trading the tech bubble for the AI bubble could be seen as equivalent exchange if we're only looking at the engine in which AI operates; spinning its wheels on itself hoping for AGI problem ignores something important
comment in response to
post
The problem with good use is that the salesman didn't sell good uses or develop because they can get more selling "AGI."
I wish more people would see that the binary between 'weak'/specific use AI and AGI is better blurred. The best solutions seem to be in between these two concepts.
comment in response to
post
I don't mean to debase the needed justifications to write about why it works in theory but there are too many of them out there and a vast number of them are undermined by the lazy AI generated writing everyone is worried about.
comment in response to
post
I'm gonna be honest: does your article actually attempt using AI in the classroom, or is this another speculative article dressed with philosophy? If it's the latter, your article is the exact type of fluff my thesis responded to. There is an abundance of "research" articles missing the empirics.
comment in response to
post
ChatGPT makes the thought of it formulaic, but I wonder if there's a set of ingredients that could be worth experimenting with much like a mixologist.
comment in response to
post
ML? I'm curious how that specifically happens since ML typically refers to NN right?
comment in response to
post
"Recent Pew Research data suggests that teachers are fully aware of the role of AI in education, with 73% of educators either seeing balanced benefits, positive potential or remaining open to exploring AI's impact"
In fact wanting to "further explore" AI impact implies the opposite of fully aware
comment in response to
post
How do you answer the regulations that limit the time in which students are allowed to take after school tutoring?
comment in response to
post
I wonder what could have motivated them to say something so heated :)
Could be Grok, highlight not related.
comment in response to
post
Though it's tempting to say "but they aren't working hard enough" or "it gives you an unfair advantage" you could make the same type of argument against tutoring or introduction templates.
The standards should not just be even-handed to the student, but should also reflect effective evaluation
comment in response to
post
The real answer from a teacher perspective is to use document tracking or pre-AI check before turning it in. In this situation, I would explain what about the supposedly AI-Generated work fails to meet the paper criteria:
- Is it too generic?
- Are they not demonstrating their thought process?
comment in response to
post
Didn't Arizona also pilot the self driving car testing to the detriment of their people? I feel like we will soon see dramatically negative effects from this.
comment in response to
post
Hi Judy! As a teacher, I've seen Khanmigo and other AI services made for teachers made publicly available so that teachers can try using them in their pedagogy. Does Google intend to do the same with Gemini or is this an option only available to Google classroom users?
comment in response to
post
To me, this is a sign that grades should be clarified & measured differently depending on the purpose. We can always make a portfolio of student work but as long as people use grades to measure talent, we will always expect teachers to simultaneously make the best and grade to out the worst
comment in response to
post
On a separate note, this quotes my thesis as concern over plagiarism and while not untrue, I wish there was more consideration for why the concern was there. Free use is one thing, but I still think designed prompts to help start things will be better in the long run. I'll code an online demo soon
comment in response to
post
Or maybe "tech-wash" is the better word, all benefits of tech aside
comment in response to
post
Who says it has to be unmeasurable? The 60 minute on Khanmigo is actually a beautiful example of trying to whitewash the social factors affecting Hobart High School and one look at the demographics and ratings after watching the video expose this.
bsky.app/profile/aewo...
comment in response to
post
TL;DR The haste and silencing of outside critical opinions instead of taking real steps to address the problem unsurprisingly caused the project to crash spectacularly. Solar laptops left out in the wind and little progress towards the goals of OCPL. Now they're careful because the money's near gone
comment in response to
post
*if the teacher is familiar
comment in response to
post
Full AI: Replicates the work env: the use of AI is graded alongside problem solving (13)
AI Exploration: Pushing the limits of AI is the task being graded (14)
This paradigm is corroborated by my qualitative study & hints at this same instinct the teacher is familiar with the tool being used 1/2