Update

These days I'm reflecting a lot on my ESL/EFL career. Much of my job was trying to teach internationals the difference between using common expressions and plagiarizing. The hardest part of this, for them, was in what was called the lit review. "Joe Smith (2000) argued that red blood cells outnumbered white ones in these conditions"....how many ways can you say this? I argued that if you actually read or knew the thing you were talking about, and someone else's lit review wasn't right in front of you while you wrote, then your report would be your own no matter how you said it. And it would be revealed to be your by not having a number of other similar sentences lifted from the same lit review appearing in your report.

While I was at SIUC its president was accused of plagiarizing large amounts of his dissertation before he graduated from that university. It is documented here deep in the bowels of the blog, but the essence of it was something like this: He was guilty; he was heading for the world of politics and eager to get out of his college experience, i.e. simply in a big hurry and not counting on needing academic validation later; and, somewhat surprised that, after a political career spanning decades, he'd come back around to his very own university, though at a tumultuous time, and be asked to lead furious faculty out of a blistering union dispute. The fact that he'd actually plagiarized ruined his career - he could never lead faculty after that. And ironically, I thought, the topic of his dissertation was gifted education.

Fast forward to today: AI is writing dissertations. Nobody can argue about whether AI was used or not, they don't have the tools to argue. AI could have been used in the construction, the revision, the drafting, the research, the constructing of the bibliography; it may not even be a big deal that AI was used. But where's the line? Perhaps the modern world is figuring out how to get what you need out of AI since we can all clearly see that it's a waste of very hard work to slog through something a computer can do for you.

Some writers are militant Luddites about the situation and some, in academia, are no doubt the same. We worked for decades to get the skills to collate information and write about it, why should we let some computer come along and take our jobs? I myself am much more neutral about it. We'll lose our jobs anyway. These things come along and take over and there's not much anyone can do.

By the way I want to introduce my new ESL site where I will try to consolidate my work. It is not, in the end, important that I publish every single thing on Amazon. More important that I know where it is, so that those looking for it can find it. Mostly it's just me, looking for it.

AI detection

There is a new trend out there. Very quickly AI has written thousands of books, and people have figured out how to use it to flood the market. Amazon tries to get authors to admit whether they have used AI. Quite a few have. They use it sometimes to fix the grammar and proofread. Sometimes they use it to do the research and organize it. Sometimes they have it write the whole thing. Why not?

There's an industry built up around "AI detection." This software will read your novel and decide whether 70% of it was written by AI, or it was 70% likely to be written by AI. Naturally this software is flawed, so writers who have written original work are dismayed to find that some programs are accusing them of having AI's help. Others have scoffed at the software, having noticed that it's somewhat random so the numbers, widely varied as they are, are meaningless. What it means essentially is that if you work hard on making your grammar standard, acceptable, and common, and have your work proofread carefully, you are more likely to be accused of AI-generated work than if you were to leave it chock full of typos.

So in my estimation it's only a matter of time before writers start leaving in typos on purpose as a kind of signature, or at the very least as a statement that a person wrote this sentence, or paragraph, or book, or whatever. How else can you tell AI from human-produced? We human authors are programmed to eliminate typos, standardize our grammar, etc. But if we need to distinguish our writing from AI writing, shouldn't we leave some typos in there? I would think so.

Here are some interesting facts. 1) if it's standard grammar, how can the software determine whether it's yours or AI-generated? It can't. It's that simple. 2) You have a style, which can be defined as the tendency to use certain constructions at certain frequencies, all else being equal, that is, assuming it's all grammatical, all prooofread and spelled properly; you still have a kind of signature in that you use em-dashes, for example, or semi-colons, at your own frequency. Ah but all this does is give AI something to work on. AI can create language at the same frequency, with your style, and still say what it wants to. So although someone can tell your writing from other writing, or AI writing, if AI is given proper instructions they can't tell your writing from AI-generated writing with your style. Both are proofread. Both have your signature and your style.

Here are some solutions: leave in a typo, or, create a typo, or create a signature typo, or, put in some bland statement about how you wrote it yourself. I'm leaning against that last one although it was my original solution.

Rather, I think we should embark on a journey of emracing and celebrating typos (although I suppose AI could generate those too, especially if they're the same in every book, every paragraph or every sentence). Let's loosen up this language and get away from the strangling tendency to FIX EVERY ERROR.

Gay & politics

Schaffer, M. (2024, Jan. 5). The Right Is Dancing on Claudine Gay’s Grave. But It Was the Center-Left That Did Her In. Politico. https://www.politico.com/news/magazine/2024/01/05/claudine-gay-resignation-battle-column-00133820.

This is an interesting article, but it's entirely from the political angle, as if the sinking of Gay had nothing to do with the plagiarism itself. It doesn't even touch the central plagiarism questions that I have always been more interested in. To me it's more a question of justice. Was she guilty? If so, should she have known she was guilty?

Personally I think left, right, ultra-left, & ultra-right all want, whether they know it or not, a disciplined scholar as President of Harvard. This person does not have to be famous or ground-breaking in his/her field. But he/she has to know what they go through on a day-to-day basis. And not to have cheated on that very same path.

Claudine Gay steps down

I am reflecting a bit about the Claudine Gay incident even though I haven't read extensively about it. Instead, I encounter her name constantly as I peddle my book Harvardinates. It reminds me of my life teaching how to avoid plagiarism and working in an institution marred by accusations against its preident (documented on this blog).

The problem is, you simply can't lead an institution of scholars when you have been proven to plagiarize. If she was guilty at all, and it appears she was, she can't just apologize, fix it, and move on, with strength and integrity behind her actions and a diverse crowd of scholars accepting her leadership. They won't. They need one of their own.

The things she plagiarized, apparently, were the hardest not to plagiarize. You've done a masssive search for material, and time has run out to wait for the library to bring you a few more things. You accept somebody's word for it that a scholar found such and such. The problem is not that scholars are wrong that some scholar found such and such. The problem is that there are only so many ways to say that. And, not having read what was written, putting it in other words puts you at risk of misrepresenting it. Trapped by circumstances, you put what you think it said on the paper. But it's remarkablyy close to what that scholar thought it said.

Looking at it from above, it's clear that the only way out is to give yourself an extra six months, and haul in every article that is even remotely connected to your topic, and painstakingly, one by one, peruse each article until the sentence you produce can be entirely your own. It's only one sentence, but it has to be entirely your own. There are ways you can use to personalize this section; the key will be that although some words or combinations of words will ding a plagiarism search, no entire sentences will, and the kind of pain that Ms. Gay has endured will be entirely avoided.

There is no question that what happened was politically motivated. Her enemies are politically motivated; so are her supporters. Everyone has other motives besides just digging through words for the pure fun of defining plagiarism in the modern day. The problem is that it is still extremely serious business, and for the scholars of Harvard, they themselves have already devoted themselves to the six-month rule outlined above in order to save their reputation and their future. Why would they not expect the same from their president? They know she has other, bigger things to worry about but so do they. They have that in common, in fact, while there is little else that they have in common.

In the 1600s a president had to be a minister. But that was very difficult, as ministers had made lifetime commitments to their congregations who then were very reluctant to let go of them. What I mean is that when Harvard needed a new President it had a hard time finding one. Finally, in about 1683, they found a guy named Nathaniel Rogers to agree to come down to Cambridge and be President of Harvard. He was a qualified minister but had been spending his time helping another one, William Hubbard, run the Ipswich Church. It took him almost a year to release himself from Ipswich and move his family down to Cambridge. But no sooner had he installed himself and his family, and took up the duties of the Presidency, in 1684, there was an eclipse of the sun, and he died right in the middle of it.

People were alarmed, but that's a different story. Life goes on. Eventually, they got Increase Mather to do it, but that had its pitfalls too. Talk about witch trials, you had to be careful in those days.

Harvard scandal

The latest on the Harvard scandal is in this Crimson article:
https://www.thecrimson.com/article/2023/12/12/allegations-plagiarism-gay-dissertation/?fbclid=IwAR1FfwIvlBOV_mnj8dacipzFS_DgvYlE-XHs9QQfOOEaFW1bALkhplOgnvc

It reminds me that as a former academic ESL teacher, I could consider myself knowledgable about what and what doesn't constitute plagiarism. What I remember is that this is definitely the trickiest part; saying what one researcher said about another can only be put in so many words. And yet as a writer you don't want to appear to not have read anything but somebody's description of what the article said.

Enough said though, since I'm retired, and this blog is full of more intelligent things said about the issue. More later.

Chat GPT

Chat GPT, if you think about it, is the greatest plagiarizer ever, besides Biden. They have actually developed a computer that studies everything that was written, learns from it, and writes just like that. In other words it is copying choice bits of sentences and putting them all together in a way that you wouldn't know it wasn't a live person.

So the question is, will the advent of Chat GPT make plagiarism more common, more acceptable, more everyday? I think so. I think we will come to recognize a "generally accepted style" and Chat GPT will be right in the center of it.

In a way, grammar correctors have already done that. They have ruthlessly gone after perfectly acceptable constructions until they have become outside the norm, outside the machine, outside the standard way. So now you have a kind of technology-crowned style. Same with Chat GPT.

And it's all because somebody somewhere wrote an acceptable story that could be copied. That's an enormous responsibility - to be a content producer in the time before Chat GPT comes to write all the content - because you know you'll be the model for generations of machine-generated language from now on.

It's awesome.

ChatGPT

From Darren Hudson Hick (Facebook),  Dec. 15:

Today, I turned in the first plagiarist I’ve caught using A.I. software to write her work, and I thought some people might be curious about the details.

The student used ChatGPT (https://chat.openai.com/chat), an advanced chatbot that produces human-like responses to user-generated prompts. Such prompts might range from “Explain the Krebs cycle” to (as in my case) “Write 500 words on Hume and the paradox of horror.”

This technology is about 3 weeks old.

ChatGPT responds in seconds with a response that looks like it was written by a human—moreover, a human with a good sense of grammar and an understanding of how essays should be structured. In my case, the first indicator that I was dealing with A.I. is that, despite the syntactic coherence of the essay, it made no sense. The essay confidently and thoroughly described Hume’s views on the paradox of horror in a way that were thoroughly wrong. It did say some true things about Hume, and it knew what the paradox of horror was, but it was just bullshitting after that. To someone who didn’t know what Hume would say about the paradox, it was perfectly readable—even compelling. To someone familiar with the material, it raised any number of flags. ChatGPT also sucks at citing, another flag. This is good news for upper-level courses in philosophy, where the material is pretty complex and obscure. But for freshman-level classes (to say nothing of assignments in other disciplines, where one might be asked to explain the dominant themes of Moby Dick, or the causes of the war in Ukraine—both prompts I tested), this is a game-changer.

ChatGPT uses a neural network, a kind of artificial intelligence that is trained on a large set of data so that it can do exactly what ChatGPT is doing. The software essentially reprograms and reprograms itself until the testers are satisfied. However, as a result, the “programmers” won’t really know what’s going on inside it: the neural network takes in a whole mess of data, where it’s added to a soup, with data points connected in any number of ways. The more it trains, the better it gets. Essentially, ChatGPT is learning, and ChatGPT is an infant. In a month, it will be smarter.

Happily, the same team who developed ChatGPT also developed a GPT Detector (https://huggingface.co/openai-detector/), which uses the same methods that ChatGPT uses to produce responses to analyze text to determine the likelihood that it was produced using GPT technology. Happily, I knew about the GPT Detector and used it to analyze samples of the student’s essay, and compared it with other student responses to the same essay prompt. The Detector spits out a likelihood that the text is “Fake” or “Real”. Any random chunk of the student’s essay came back around 99.9% Fake, versus any random chunk of any other student’s writing, which would come around 99.9% Real. This gave me some confidence in my hypothesis. The problem is that, unlike plagiarism detecting software like TurnItIn, the GPT Detector can’t point at something on the Internet that one might use to independently verify plagiarism. The first problem is that ChatGPT doesn’t search the Internet—if the data isn’t in its training data, it has no access to it. The second problem is that what ChatGPT uses is the soup of data in its neural network, and there’s no way to check how it produces its answers. Again: its “programmers” don’t know how it comes up with any given response. As such, it’s hard to treat the “99.9% Fake” determination of the GPT Detector as definitive: there’s no way to know how it came up with that result.

For the moment, there are some saving graces. Although every time you prompt ChatGPT, it will give at least a slightly different answer, I’ve noticed some consistencies in how it structures essays. In future, that will be enough to raise further flags for me. But, again, ChatGPT is still learning, so it may well get better. Remember: it’s about 3 weeks old, and it’s designed to learn.

Administrations are going to have to develop standards for dealing with these kinds of cases, and they’re going to have to do it FAST. In my case, the student admitted to using ChatGPT, but if she hadn’t, I can’t say whether all of this would have been enough evidence. This is too new. But it’s going to catch on. It would have taken my student about 5 minutes to write this essay using ChatGPT. Expect a flood, people, not a trickle. In future, I expect I’m going to institute a policy stating that if I believe material submitted by a student was produced by A.I., I will throw it out and give the student an impromptu oral exam on the same material. Until my school develops some standard for dealing with this sort of thing, it’s the only path I can think of.