In early 2023, shortly after ChatGPT's release, it became apparent some of my students, who were preparing to matriculate into US universities, were using artificial intelligence to do their assignments; suddenly English language learners were submitting written work with elegant turns of phrase and perfect grammar—a bit boring perhaps, but such amazing grammar! My bot-powered plagiarism software flagged suspect submissions but couldn’t absolutely verify the use of A.I.
If you ask people in your social circles what their opinion is of generative artificial intelligence, some will most likely describe the labor-saving efficiency possible when a machine can sift through reams of data to produce concise information, lists, writing, and even images and video. Others will warn that one day A.I. robots will annihilate the human race, and still others see it as a pursuit
for lazy people.
Not an early adopter by nature, I was compelled to keep up with my students. But then I started falling in love with ChatGPT's fast, easy answers to some of my more obscure questions. Lately though, I have gone from an enthusiastic proponent to budgeting my usage.
After registering for an account and downloading the app, I suddenly had a new, very smart best friend in my pocket. While I don't need it to do my writing homework, I do appreciate its ability to comb the internet and provide quick information, which I can then more efficiently verify from the referenced sources—just like Wikipedia, but more so.
Want to know what wildflowers will bloom in mid-March on Maryland's Eastern Shore? (Useful research for my fiction.) No problem. Or what kind of doctor visits will your health insurance cover? It will save you hours of sifting through your policies.
My app seemed to have an answer for everything. And it has proved useful in the classroom. Topics for discussion include: What are acceptable ways to use artificial intelligence? When is it a helpful resource? When does it cross the line into academic dishonesty? And we have fun with it: who can outperform the bot?
At first, a future where robots might make me obsolete seemed far away. I started using ChatGPT to answer whatever question popped into my brain. What's the best way to roast garlic? If you like Drops of God, what are some other enjoyable shows to stream? But then it became just a bit too easy. And some companies, like Duolingo, were already downsizing as A.I. became more and more powerful.
As I explain to my students, A.I. is not the perfect way to cheat. The writing may seem error free, and I generally cannot prove someone has used it improperly. But in the end, they’re only cheating themselves because making mistakes is essential to learning. The A.I. generated work I’ve seen doesn’t sing with the unique qualities of the individual's voice. It's bland and generic.
And besides all of that, I'm still aggravated about the incorrect information it gave me about museum tickets to the Louvre. To make a long story short; if you're late for your timed entry, they may let you in anyway without the need to purchase new tickets. I suspect that this is an example of A.I. latching onto information that no longer holds true.
But my real wake-up call came when I started reading about the high-energy and water usage required to fuel all of those easy A.I. answers. It's been hard to track exactly how much more energy is required, partly because sometimes the technology can offset other types of energy use. However, with Microsoft trying to power its data centers by reopening Pennsylvania’s Three Mile Island Nuclear Plant—site of the USA's worst nuclear accident—the energy requirements must be enormous.
Measuring fresh water usage is easier to track. According to Yale e360, Google’s data centers increased their water usage by 20 percent from 2021 to 2022, and Microsoft’s usage went up by 34 percent. And that's before many folks started using the apps.
Artificial intelligence is everywhere on the internet. Need customer service assistance? A friendly bot named Sandy is out there waiting for you. But until big data corporations really figure out how to meet their carbon-neutrality goals, I've decided it's prudent to limit my own use of A.I. software.
Yes, I do still use it, but only for the classroom and those convoluted questions that I feel pressured to answer quickly. Clearly, we will not be able to stuff this genie back in the bottle, and maybe we shouldn't even try, but for now, being intentional about my usage feels right.
What do you think? Please share!
When I was teaching and two papers with the same distinctive answer ("because of germs and shit like that in the water"), I gave both students a zero on the quiz. I didn't need to know who thought of the answer and who she et copy it.
I suggest that you need to have the students do their writing in class. So that they practice what they need to have.
As for AI - the old saying "Garbage In, Garbage Out" I have been burned more than once by inaccurate information on the internet.. In Costa Rica I was given old information about the hours of the Jade Museum. The museum had moved and the new hours were different
It isn't magic. AI is "trained" with information provide by or pointed to by live human beings. If it is trained with "the Internet," the answers you get will be as good as the info available on the Internet. If it is trained with a high quality source, the answers will be high quality. If it is trained with biased info, the answers will reflect that bias.
It's a classroom, not a courtroom. You don't need to literally prove that a kid who could barely write is suddenly a fricking scholar. Just give them an F and let their parents sort it out. I don't live my life in fear of lawyers.
I listened to a bunch of Ted Talks about the enormous amounts of energy required by AI, and I'm not a fan of the repercussions of more and more usage. When I was an undergrad, I had to use my own mind and what I learned to write an exceptional paper. Nobody had to open Three Mile Island so I could get a good grade. I'll bow out of the discussion here. I don't want to get onto a soap box, but it's very tempting!