In the Dark<p><strong>Generative AI in Physics?</strong></p><p>As a new academic year approaches we are thinking about updating our rules for the use of <a href="https://en.wikipedia.org/wiki/Generative_artificial_intelligence" rel="nofollow noopener" target="_blank">Generative AI</a> by physics students. The use of GenAI for writing essays, etc, has been a preoccupation for many academic teachers. Of course in Physics we ask our students to write reports and dissertations, but my interest in what we should do about the more mathematical and/or computational types of work. A few years ago I looked at how well ChatGPT could do our coursework assignments, especially Computational Physics, and it was hopeless. Now it’s much better, though still by no means flawless, and now there are also many other variants on the table. </p><p>The basic issue here relates to something that I have mentioned many times on this blog, which is the fact that modern universities place too much emphasis on assessment and not enough on genuine learning. Students may use GenAI to pass assessments, but if they do so they don’t learn as much as they would had they done the working out for themselves. In the jargon, the assessments are meant to be <em>formative</em> rather than purely <em>summative</em>.</p><p>There is a school of thought that has the opinion that formative assessments should not gain credit at all in the era of GenAI since “cheating” is likely to be widespread. The only secure method of assessment is through invigilated written examinations. Students will be up in arms if we cancel all the continuous assessment (CA), but a system based on 100% written examinations is one with which those of us of a certain age are very familiar.</p><p>Currently, most of our modules in theoretical physics in Maynooth involve 20% coursework and 80% unseen written examination. That is enough credit to ensure most students actually do the assignments, but the real purpose is that the students learn how to solve the sort of problems that might come up in the examination. A student who gets ChatGPT to do their coursework for them might get 20%, but they won’t know enough to pass the examination. More importantly they won’t have learnt anything. The learning is in the doing. It is the same for mathematical work as it is in a writing task; the student is supposed to <em>think</em> about the subject not just produce an essay.</p><p>Another set of issues arises with computational and numerical work. I’m currently teaching Computational Physics, so am particularly interested in what rules we might adopt for that subject. A default position favoured by some is that students should not use GenAI at all. I think that would be silly. Graduates will definitely be using CoPilot or equivalent if they write code in the world outside university so we should teach them how to use it properly and effectively.</p><p>In particular, such methods usually produce a plausible answer, but how can a student be sure it is correct? It seems to me that we should place an emphasis on what steps a student has taken to check an answer, which of course they should do whether they used GenAI or did it themselves. If it’s a piece of code to do a numerical integration of a differential equation, for example, the student should test it using known analytic solutions to check it gets them right. If it’s the answer to a mathematical problem, one can check whether it does indeed solve the original equation (with the appropriate boundary conditions).</p><p>Anyway, my reason for writing this piece is to see if anyone out there reading this blog has any advice to share, or even a link to their own Department’s policy on the use of GenAI in physics for me to <s>copy</s> adapt for use in Maynooth! My backup plan is to ask ChatGPT to generate an appropriate policy…</p><p><a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://telescoper.blog/tag/assessment/" target="_blank">#assessment</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://telescoper.blog/tag/chatgpt/" target="_blank">#ChatGPT</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://telescoper.blog/tag/copilot/" target="_blank">#Copilot</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://telescoper.blog/tag/education/" target="_blank">#education</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://telescoper.blog/tag/formative-assessment/" target="_blank">#FormativeAssessment</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://telescoper.blog/tag/genai/" target="_blank">#GenAI</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://telescoper.blog/tag/generative-ai/" target="_blank">#generativeAI</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://telescoper.blog/tag/physics/" target="_blank">#Physics</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://telescoper.blog/tag/summative-assessment/" target="_blank">#SummativeAssessment</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://telescoper.blog/tag/theoretical-physics/" target="_blank">#theoreticalPhysics</a></p>
#Copilot
8 posts7 participants0 posts today
Continued thread
Right... so I asked 'Support' about it (probably another LLM) and it says that the solution is... that it can't be disabled ?!?!
Am I becoming crazy? The Copilot toggle is OFF in this screenshot, right? If so, why is there still a Copilot icon on the top right? How do I get rid of it??
(this is in #Outlook for Win10)
Edit: the (temporary, because I already did it once a few days ago and Copilot is now back) solution is to follow the instructions for editing group policies here: https://pureinfotech.com/disable-windows-10-copilot/ (Method 2 for Win 10).
It's "the year of AI agents". Also, 2025 is yet another year when M365 Copilot cannot yet turn an email into a calendar entry inside Outlook.
Sorry, that was too harsh. It CAN do that -after these 10 steps that I went through. Which included lots of failures and gashlighting, LLM style.
Why everyone isn't paying $30pupm for #Microsoft365 #Copilot at this point yet is a real mystery.
Longer thread on LinkedIn: https://www.linkedin.com/feed/update/urn:li:share:7355896589372030976/
This is really wild: Microsoft is pushing more and more #spyware features into everyday products to steal all of your data:
"#Copilot #Vision is an extension of Microsoft's divisive Recall [...] is designed to analyze everything you do on your computer [...] by capturing constant screenshots and feeding them to an optical character recognition system and a large language model for analysis – but where #Recall works locally, #CopilotVision sends the data off to #Microsoft servers." 
https://www.theregister.com/2024/05/28/microsofts_recall_preview_on_non_ai_pc/
I would never ever trust a Windows system not to send out anything I do.
The Register · Microsoft's Recall preview doesn't need a Copilot+ PC to run
Replied in thread
@carnage4life And I’d always take any company’s research into its own products with a massive, Everest-sized grain of salt. When I worked at Amazon, there’s zero chance they’d let anything out if it was at all unfavorable. #AI #CoPilot #LLM
"AI" apologists are now touting fucking blockchains as a solution to making sure artists' rights are respected when training models and generating slop.
This industry should die already.
Replied in thread
@jwildeboer wow #copilot #ai you are using as a corporate #microsoft user is running in a #jupyter notebook container. These #infosec researchers hacked it to find out what's what and so on. Interesting read.
#linux and #python rule for sure
Explanation on jupyter here: https://docs.jupyter.org/en/latest/what_is_jupyter.html