It's me. I'm the gunner.

So last year, I either 1) accidentally got accepted to a graduate program, or 2) my credentials were so amazing that they admitted me with an incomplete application. Frankly, I don't know which it is at this point. I was convinced they had made a mistake, but after interacting with some of my classmates, I am thinking it might actually be the latter.

One of my courses this semester is a qualitative research course. I'm not new to qualitative research - one of my first major-output projects was a mixed methods study, and I was responsible for transcribing interviews. By hand. Because automatic transcription didn't exist back then (or at least not without paying $$$ for software). I painstakingly put audio files on 0.5x and typed and retyped and revised and edited.

Our module this week is on transcription and coding. One of the assignments was to transcribe an interview video on YouTube and share our results on a discussion board to, well, discuss. Easy enough, right? How do you mess that up?

One of my classmates fed the interview to an LLM, most likely ChatGPT (based on phrasing, punctuation, and hallucination patterns). Her resulting transcript was full of horrendous holes. Phrases that were edited just enough to not mean the same thing. Phrases omitted completely. Statements misattributed to different people. Terminology being given abbreviations in parentheses when those abbreviations were never spoken aloud. Some phrases were straight up hallucinated. In the middle of one of the speaker's long rants, ChatGPT inserted this sentence, which it stated was from the interviewer: "What is a Practice Research Network?" The interviewer had not uttered these words. He hadn't said a thing during the course of that entire rant by the interviewee.

Normally, I would just let it slide. I had seen classmates use ChatGPT for their discussion board questions before and just shrugged it off. Some of them were ESL speakers who may have used AI to help them edit. I didn't have clear proof (beyond the very obvious speech patterns of ChatGPT, but that's not watertight proof); so I simply haven't said anything until now.

Not tonight, though. Tonight, after an annoying day at work wherein I had to teach therapists how to do their job, after I engaged in supervision sessions with several individuals who had no idea what they were doing, after I had to handhold a research associate through a simple task... tonight I had had enough.

Not only did I call her out, I came with receipts. Because good research involves proof 😤

Another one of my classmates copy/pasted YouTube's auto-generated transcription. If you've been on YouTube at all, you know that those transcriptions are horrid. They make no sense grammatically; they are full of spelling errors; there is no capitalization or punctuation... and they have awkward line breaks. My classmate tried to at least minimize the line breaks on the first page, but she just flat out gave up on the second page onwards. Below is the Youtube transcript on the left, my classmate's work on the right:

So naturally, she got a response too (the formatting was unfortunately messed up in the process):

I'm not usually a tattletale and I'm not usually a gunner. But more and more I'm realizing that if we don't do the work of gatekeeping and policing our own, no one else is going to do it. This is a doctoral program, mind you. These are students who will one day write their own dissertation, publish the research, go out and see patients. This is unacceptable.

I'm not anti-AI. I've used Claude to help sort through my messy brain and tease out a research topic that is feasible. I've asked ChatGPT to look through my R code and help me figure out why the heck I was getting p-values of 0.98. I've used Elicit to extract pooled effect sizes (a singular metric that can be hard to hunt for) from a pile of 100 papers to identify research gaps. They're tools to help us do work, not to do the work for us.

And at the end of the day, it's not even the AI usage I'm frustrated by.

It's the fact that if you're a doctoral student, who didn't even think to TRY to mask the fact that you used AI, who just copy/pasted it verbatim thinking no one else would notice... if you're that foolish, then you don't deserve to be issued a PhD. Go home.