I think a fundamental problem in my understanding of your main point is that you are not expressing it clearly. In particular, you seem to be using the term “write” at times to mean:
“The ability to (grammatically) express one’s understanding of a topic through the use of proper English,”
and at other times to mean:
“To have a level of understanding necessary to produce a (logically) coherent document on the subject.”
Let’s go through your posts in this thread to see what I mean:
@edyvw :
“Any hint.” That appears to cover both meanings of the term “write.” Inclusive, thus ambiguous, thus it doesn’t clarify.
@edyvw :
Generically using the term “write,” so still no clarification.
@edyvw :
This seems clear—students are getting into trouble with AI because they don’t know the hows and whys of the topic. So, you appear to be using the term “write” in a broader sense to include the ability to collect and understand info on the topic to be written about.
Got it.
@edyvw :
Although less clear, I continue to think you are using “write” to mean “understand,” but it’s ambiguous as it’s possible that the student simply didn’t want to do all that writing after doing all the reading to (successfully) come to a full understanding.
@edyvw :
Yes, but what did the student (or you) mean by “write?” Understanding the material, or just taking the time to get words on paper after full understanding?
@edyvw :
Ok, this seems to provide good clarity—you're using the term “write” more generally to include understanding.
@edyvw :
More clarity as you repeat that using AI is “fine” after understanding the material. And, since you (below) have clearly defined “using AI” as
not including grammar-checking tools, apparently using ChatGPT after understanding is “fine.”
@edyvw :
This reinforces the previous clarifications as you are saying that using grammar-checkers (which are often AI-free) is ok but using it to write the whole paper is not. So, “write” clearly means understanding the material.
Starting to feel like I’m getting what you’re trying to say.
If this is indeed what you are saying, I would ask this: you seem to be working under the premise (axiom, really) that having ChatGPT write a whole paper directly implies the student will not know the hows and whys.
While I understand that this will many times be true (as the weaker students will often lean on AI to make up for their deficiencies), is it not possible that a very strong student (or even you) could save a ton of time by prompting ChatGPT what they want written based on their stone-cold solid understanding of the material?
Thus, creating a paper that might even synergize with the student’s (or your) expertise to make the paper even better?
Could there be advantages for people who are strong in the subject to use this approach rather than bothering “colleagues, spouse, girlfriend, boyfriend, mistress, lover, brother, sister, co-worker, whoever” to provide the same feedback (but likely at an inferior level)?
You seem to be saying that the difference between 1) using a grammar checker + multiple human reviewers and 2) using AI to do the same thing is a rationale to ban AI use by students. Does that supposed difference really support such a strong and decisive conclusion on AI use in the classroom?
At least from what you have presented, your case is tenuous at best and needs far better support than you have provided.
@edyvw :
Ok, we’re back to ambiguity: after running the work through a grammar checker, you encourage the work to be then sent to a litany of potential human reviewers for “feedback.”
What is this feedback? A second (or multiple) check(s) on grammar? A review of the substance of the written work? If the latter, how is that different from using AI?
@edyvw :
Ok, earlier clarity negated and I’m lost again. The most direct reading of that is that you are back to using the term “write” to indicate grammar and flow, not understanding. If so, this is inconsistent with earlier clarifications that indicate the opposite. So, I’m no longer sure of your point.
@edyvw :
Again, not clear because you are not defining your terms well. The first sentence indicates your beef with AI is in using it to create the substance of the paper, but the last sentence indicates that you have reservations about using AI for grammar correction.
You need to be clearer on exactly what your problem with AI is.
@edyvw :
You are now bringing up a question about AI ethics, which is a very broad topic, certainly not limited to using AI to “write.”
Your first comment is about ethics related to a suicide case, which makes it unclear as to how you’re trying to connect ethics to whether AI should be banned for students.
Though I can probably connect the dots as to the connection you’re trying to make, it shouldn’t be on me to figure out your point. Worse, this feels more like a tangent that only weakly supports your conclusion and feels more like a diversion rather than an attempt at support.
@edyvw :
This summary is not even close to supporting the point that (I think) you’re trying to make, and in fact makes the opposite point.
While the expansion of literacy among the elite and skilled workers was a factor (among many) in post-Dark Age industrial development, literacy among the general populace significantly lagged the industrial revolution. Additionally, books were largely unavailable to the general population due to cost.
I believe there is evidence that many of the countries that topped the industrial revolution race had among the lowest general literacy. I’m not an expert in this (I suppose I could consult AI!), but on its face, your point (and your summary to support it) is completely vacuous.
You said you didn’t want to explain this connection to the student as it was a “fool’s errand.” Based on what you just wrote, I think I tend to agree, but with a very different interpretation of where the foolishness in that explanation would lie.
Perhaps you could flesh this out a bit more (though I'm not certain it's worth it-- the point seems terminally flawed).
@edyvw :
I would say that, due to very poor writing skills + poor understanding of the fundamentals of the topic, you have utterly failed to support that conclusion (or even make clear what point that conclusion was supposed to support).
@edyvw :
I do.
I was a math professor for 20 years, nearly half of that as department chair. 34 faculty as direct reports, numerous hires, and numerous promotion and tenure decisions.
I’m retired now, but I was well-funded (externally) and well-published. Two major works published in the Proceedings of the National Academy of Sciences.
Even though I was head of a math department, my Ph.D. is actually in Biochemistry/Cell Biology (I’m a mathematical biochemist), so I also served many years on the college promotion and tenure committee, reviewing the research of faculty ranging from Black history to particle physics.
I think I know something about academic writing, and I definitely know quite a bit about faculty (having dealt with them extensively over many years as a department chair). Despite often being well-versed and even brilliant in their fields, faculty are often:
1) Surprisingly bad writers—and I mean both in grammar/flow, and also in the logical structuring of their writing
2) They uniformly believe their writing skills are strong, even when clear evidence to the contrary is presented (which I have done many times).
While not declaring you objectively “wrong” on substance, I have found your points to be neither well-formed nor well-supported, and I feel bad that students are forced to live under your “no hint of AI” rule as that is going to be a huge disservice to many of them and you don’t seem to have a cogent rationale to support doing that harm to them.
You may not be interested, but here’s my take:
Weak student + no AI = poor performance
Weak student + AI = poor performance, but marginally improved (for a number of reasons). But still not good
Strong student + no AI = strong performance
Strong student + AI = emergently strong performance, where the sum of the individual performances is more than the parts
Thus, I would conclude that your arbitrary and capricious AI rules are almost certainly creating damage to strong students, and possibly damage to weak students as well. And rather than presenting a rationale for that damage, you’re mostly just going on an old-man rant about how “that’s not now I did it, and by golly, my students aren’t going to do it either!”
My support (in part) for this take?
@KrisZ just posted a link to a paper that presents correlational effect of AI on weak students (with causality and its direction unaddressed), but also clearly echos my framing for the stronger students as it cites a study that (emphasis added):
More importantly, it says that:
Hardly support for "any hint of AI and you will be thrown out of the university!!"
And it nicely states my perspective on the topic: the effects of AI are almost certainly nonlinear, thus almost certainly mathematically “complex” (defined as having both an ordered and chaotic element), and therefore not reducible to simple conclusions such as knee-jerk bans of its use by students.
Though I have taught (very) basic neural network and learning model theory, that treatment was very superficial, so I am decidedly not an expert in AI.
However, I think I do know enough to be able to say with certainty that AI effects on critical thinking is a complex topic and making binary, blanket statements is not a good way to approach the topic.
For the sake of your students, I would plead with you to take a more nuanced approach to AI in your classroom. Do the deep dives you expect of your students, steelman the pro-AI arguments, and strive to achieve balance and nuance in your conclusions.
