Post by Morgan Robertson / Redsky

Morgan Robertson • Feed

Important reading as you prepare for your semester

aug 31, 2025, 12:10 am • 364 103

Replies

This definitely helped me with my course statement on AI and large language models!

The use of large language models such as Chat GPT caa be a form of academic misconduct, covered by a number of the categories outlined by the University. In particular, it involves presenting work as ones own when that is not the case, and so is a form of cheating. If that were not reason enough, there are a couple of other things to consider. First of all, large language models are predictive and are not intelligent. They operate on the likelihood of one word following another. This is not designed to produce reliable statements of truth, nor arguments based on an analysis of evidence and an application of evidence. Rather, as observed by Hicks, Humphries and Slater (2024), they produce something that is designed to appear plausible, but has no concern with truth. It is, as they say, bullshit. Secondly, you are here to learn and develop your critical thinking skills and your communication skills. The process of writing, and of receiving feedback, is fundamental to that. It is an exercise. You are not being asked to write papers because of a need for more papers to be produced in the world (thank you to Prof Emily Bender of the Mystery AI Hype Theater for this point). Literature lecturer Carrie Shanafelt recently observed that you would not take a forklift into the gym in order to lift weights. The wieghts do not need to be lifted, rather you need to lift the weights to maintain and develop muscles. So, in summary - you don't want to be accused of cheating (and penalised for it), you don't want to submit bullshit in place of something that reflects your own real intelligence, and you don't want to throw away the chance for developing through your study programme. References: Hicks, M.T., Humphries, J. & Slater, J. (2024). 'ChatGPT is bullshit'. Ethics and Information Technology 26(38). https://doi-org.eux.idm.oclc.org/10.1007/s10676-024-09775-5 Points from Bender and Shanafelt can be found in their postings on Mastodon.

aug 31, 2025, 11:49 am • 13 7 • view

Don’t start me.

aug 31, 2025, 1:03 pm • 1 0 • view

📌

aug 31, 2025, 3:08 pm • 0 0 • view

They were hosted on the Ed Zitron podcast. What a lively bunch. Great episode (aug. 13 2025)

aug 31, 2025, 10:15 am • 0 0 • view

I love this paper! It was very helpful as I was wrapping my head around the ethical problems surrounding "AI".

aug 31, 2025, 6:37 am • 2 0 • view

"Calling these inaccuracies ‘bullshit’ rather than ‘hallucinations’ isn’t just more accurate (as we’ve argued); it’s good science and technology communication in an area that sorely needs it."

aug 31, 2025, 12:59 pm • 3 0 • view

link.springer.com/article/10.1...

aug 31, 2025, 12:10 am • 19 2 • view

aug 31, 2025, 12:35 pm • 4 1 • view

I think the major concern w ChatGPT and AI in general is that people are breezily unconcerned as to the truthfulness & accuracy. It is known that it makes up case law & lawyers still use it. People do not care anymore. We are at the fatal laziness stage.

aug 31, 2025, 10:17 am • 4 0 • view