May 4, 2023

[The AI Hype Wall of Shame is a collaboration between Critical AI and the DAIR Institute. For more on this initiative and our rating system, see this link.]

Some of the best reporting on “AI” comes from @techreview, so we regretfully unfold a new entry for #AIHypeWallofShame: a 4/6/2023 article by Will Heaven that purports to be about education.

Headlines are often written to grab attention–chosen by editors rather than journalists themselves. But here’s a case in which the provocative headline (“going to change… not destroy”) is less accurate than the subheading (“cheating doesn’t tell the whole story”). 

The truth is that no one has any idea what “ChatGPT” will do to “education,” though we can be sure that OpenAI is glad to get another headline. This article, alas, will tell you very little about the nominal subject. 

On our AI Hype scale, we give it 1 Marvin (boring). 

The article begins by making a perfectly reasonable claim by Jenna Lyle (Deputy Press Secretary for the New York City Board of Education), its baseline for the alleged “initial panic” surrounding large language models (LLMs).

Can we charge $100 whenever publications use “panic” to describe what teachers say or do in response to a new technology? Where, after all, is the evidence that ChatGPT does “build critical-thinking and problem-solving skills.” Hint: not anywhere in this article.

We’re next told that three months on, “the outlook is a lot less bleak”–here again presuming that Lyle’s position is or was an emotive state of “bleakness” as opposed to a perfectly reasonable statement about student learning criteria.

But let’s be fair. Heaven claims to have interviewed teachers. Let’s hear from them! Heaven tells us that…

Notice how no actual teachers come up—only ed-tech companies and OpenAI which (while commercializing subscriptions to ChatGPT with a $10 billion-plus  investment from Microsoft) has plenty of, um, incentive to persuade educators that their product is best thing in education since the chalkboard. 

So for sure, the very next quotation will be from a teacher, right?

Ok, so no teachers yet, but we can be even more certain OpenAI is “engaging with educators.” As a reader, you might wonder about the purity of the motives behind this “engagement.” You might also wonder about the source of OpenAI’s newfound experience in educating students.

Two more paragraphs in and we STILL have nothing from an actual teacher. It will take time for “overworked” educators to “innovate.” And banning is “futile” 

(note that last time I looked you needed to be 18 to subscribe to OpenAI which is something that suggests that even OAI is not certain that it’s product is ready for schoolkids?).

It may be “far too soon to say what the lasting impact of ChatGPT will be,” but apparently it’s never too soon for a journalist to rehearse industry talking points while claiming to speak for teachers. Should we understand now that teachers are simply too “overworked” and “under-resourced” to have opinions that might be worth printing in an article ostensibly about what teachers think?

Reading that paragraph it’s impossible to know who Heaven is interpolating about “feel[ing] bigger” and “one way or another.”  

Himself? 

An ed-tech CEO? 

Is there no teacher able and willing to add to this story? 

As we break to a new section, the only “teacher” quoted is the one who allegedly “panicked” back in November.

But did they “freak out”? 

Who? 

NOTE: If you are a teacher and “freaked out” in November and again in March, please write to @techreview so they can substantiate their reporting of your mental state. 

Heaven continues,

Again, whose opinion is that? 

Heaven’s? Or these mystery teachers he interviewed?

Here (from a different article) is another way of looking at NYC educators that doesn’t reduce them either to panicked deer in the headlights or dumb-dumbs in the shadow of some really smart software: 

A few sentences down, a teacher (!)–actually a UK bioscience professor, but okay–testifies that his students hadn’t heard about ChatGPT. 

So a “tempest in a teacup.”  “Even so,” he says, teachers are right to see the technology as a game changer.” LLMs “are set to have a massive impact on the world.” 

Wow, it’s amazing how tempests in teapots because massive game changers in the wink of an eye! 

Can we be clear that 

1. 1. lots of things are being rolled out into consumer and business software that don’t need be taught in a classroom;
   2. there is indeed a key distinction between teaching students about how a technology works and deciding that one must teach with that technology. 

Heaven quotes an administrator at the University of Baltimore:

In principle I agree about not getting stuck in fear: but who exactly was frightened in the first place? 

Deep into the article, without a single example of a discussion with a teacher about how chatbots might be used well in the classroom, we get a vague survey: 

If you’ve never heard of the Walton Family Foundation, as you’ll guess, their philanthropy is tied to Walmart and one of their three areas is “strengthening the connections between K-12 education and lifelong opportunity.” 

Is there no other way for MIT Review to hear from teachers?

Well we hope so too.

Now let’s see some evidence in an article claiming that this technology will “change” education.

In what’s below, a professor (note that K-12 teachers only get to “panic” in this article) tries out an actual idea–one we’ve seen a fair bit of since last November.

It seems reasonable. 

Certainly that’s one approach to how to teach “critical thinking” through writing. Of course, it’s been criticized for a number of obvious reasons, but surely we’ll hear about the strengths and limitations of this single approach? 

Alas, no—instead we segue to the UK bioscience professor now calling out his colleagues for “hand-wringing” of the kind that reminds him of the pandemic.

Darn those panicked, fearful, hand-wringing educators!

(I feel so superior to them, don’t you? I spent the pandemic thinking I’d gone to heaven.)

Here’s the wrap, folks. 

We apparently have evidence for the benefits of “game-changing” and “massive” impact of a technology that’s been around for less than a semester: courtesy of OpenAI, two ed-tech companies, two happy professors, an administrator, and a survey from the Walmart family. 

You heard it here first—there’s something called the INTERNET, and information is now “everywhere.” 

In case you missed the above while hyperventilating, even when bots misinform, it’s an opportunity for teaching! 

It’s changing everything, because like, it’s there, and, you know, when it’s perfect that’s great, and when it’s not perfect that’s even more great, because…teaching. I cannot lose so long as my students subscribe NOW (whether they’re over 18 or not?) to learn with for-profit chatbots.  

It WILL build critical thinking skills because, like, it’s “new,” it’s “online” and…

I’m not hand-wringing, or tempest-cupping.

I’m cool, I’m dope!

Please Mr. Heaven, the next time you promise to report on a “number of teachers,” can you do more than fan service for your preferred narrative? 

Don’t gaslight teachers in the act of supposedly reporting their views. They are, after all, the domain experts on this topic.

Thanks for reading.