FONK, not FOMO

I’m not going to rehash here the myriad debates about what Artificial Intelligence (AI) will do to education and colleges. There’s been enough electrons and ink spilled about that, with more to come. I want to focus on one aspect that ties in well with the typical academic’s tendency toward imposter syndrome — the fear of not knowing. There are two aspects I want to consider for a moment: the fear of not knowing whether something was written using AI, and perhaps more interestingly, the psychological need to know more and more about subjects in an attempt to discern whether something was written by AI.

The first fear is ever-present for those of us in education — we don’t want to give the “A” to something completely generated by AI. Then again, we’ve given “A’s” to papers written by someone other than the original student before — whether it be through paper mills, copying assignments from previous semesters, and so on. We’re not 100% foolproof, but yet we want to be. That’s a given. That won’t change with AI, given that AI checkers themselves can be fooled^[1]No, I’m not going to tell you how…there are plenty of sites out there that claim to provide such direction, and given that those sites will likely change over the next few years, to list … Continue reading. Sites such as Grammarly have tripped up some AI detection.

It’s the second aspect that I find of interest, because I don’t think it’s been talked about very much. We know that various tools such as ChatGPT and similar tools don’t always utilize correct information^[2]Oh, Reddit. Such very good and very bad information you once had…. The tools have been known to make up sources, including newspaper articles. This is the part that makes me a bit worried. I’ve been known to go through various assignments and double-check the sources (my Argumentation and Debate students know of many examples where I pulled evidence to help their cases that they found – not me – but didn’t use in their briefs). But to check every single paper and follow every single source? That would take a mind-blowing amount of time. Most of the time, it’s not as simple as (not) finding a journal called The Whale in a paper about Moby Dick. ^[3]That one actually happened in a high school class. One of my very bright colleagues wanted to test whether our teacher actually read the papers.

That’s why I think of my role as a professor in this AI world with perhaps a bit more trepidation — as someone who needs to know more than ever before… or at the very least, know how to research and learn more than ever before. I feel FONK — the Fear of Not Knowing. In my case, it’s the fear of not knowing whether something’s made up or not if a student turns in a paper that uses generative AI. And perhaps, on a more existential level, while we know that not everything on the internet is true^[4]I know, shock, right?!, the more that “credible” sources rely on the admittedly early versions of generative AI that we have now, the more that their credibility will potentially decrease (see CNET, Amnesty International, etc.). It’s pretty easy at this point to tell the sports stories that use generative AI when reporters aren’t there — that kind of service has been around since 2018, and arguably before. But as AI gets more sophisticated, it’s going to be harder to figure out what’s true and what isn’t. And that means I feel like I need to know more in order to know whether or not what I’m reading has any validity. Perhaps it brings new meaning to “doom reading.”^[5]I use this term instead of doomscrolling, because it may mean we go back to <gasp> books!

There’s a whole set of implications for generative AI and political communication — but I’ll save that for another post.