It was late December when I first noticed articles about ChatGPT.
I was fascinated, and as an ‘80s girl and a fan of the movie “War Games” (Shall We Play A Game??), I was also slightly concerned about what this type of generative artificial intelligence could really do.
But I figured, “Hey, we have Siri, we have Google that fills in your search, we have chatbots that help on websites. How different could this be?”
When I say this will blow your mind when you first encounter it, I mean it. I started experimenting with ChatGPT in January on my newsroom computer. Anyone can access it, so I created a username and login -- easy as that. I asked it to write television news scripts on everything from the U.S. debt ceiling to a documentary on the old Hudson’s building in downtown Detroit.
It spit out accurate text rapidly, within seconds. Then I moved onto a three-paragraph paper about St. Joan of Arc written at an 8th-grade level. It delivered in 10 seconds. A sonnet about cowboys? Yep. A grocery list for fettuccine alfredo? An explainer on why ChatGPT wouldn’t take my job as a journalist?
The chatbot explained that it is only as good as its programming and could never replace the judgment of a human. I wasn’t totally reassured.
If I was messing around with ChatGPT for fun, I could only imagine what people who were, ahem, younger and more tech-savvy were doing. Yep, college students have found this, my high school student has found this, and they’re using ChatGPT’s capabilities for bigger things than grocery lists and sonnets. Coding, calculus help, outlines on symbolism found in classic novels. The list goes on, and they’re not alone.
As I learned more and spoke with computer science experts, it’s become very clear that this is just the beginning of tasking generative AI with more complicated jobs than we thought. And schools have to be very aware of what tools kids are starting to use. But I don’t think banning this technology is the answer (though it’s already banned at NYC public schools, LA, and Seattle).
Regulation should be a priority, as well as more software that can detect computer-generated text. Though ChatGPT is learning every time it interfaces with a user. It has already been attached to Microsoft’s Bing search engine (Microsoft is a multibillion-dollar investor in OpenAI). Google did a soft launch of its AI a few weeks ago. And Meta has one in the works too.
(PS if ChatGPT could actually do my grocery shopping for fettuccine alfredo and then make dinner, then we’d be talking revolutionarily.)
Click here to check out ChatGPT for yourself.