Understanding ChatGPT's capabilities
What if I told you that it took me 20 seconds to write this story? Well, that would be nonsense. I asked ChatGPT to write a story about itself and it did that.
Machine Learning Scientist for Tunrnitin.com Zack Bennett said you can ask it to do just about anything.
“It allows you to interact with something called a language model,” he said. “And a language model is something that’s been trained to look at all kinds of human writing, like tons of web pages. Digital books. Wikipedia. And after it’s done training, when you go to the Chat GPT website, you type in something called a ‘prompt.’ Which is a command. So, you can ask it a question and ask it to do a task for you. And it will generate text that matches what you asked it to do. So, if you go to it and ask it to write a 3-page essay about the novel ‘Moby Dick,’ it will do that for you. It will generate a 3-page essay for you in a matter of seconds.”
Its abilities to spit out essays, brief complicated legal memorandums or even write code are impressive. I even asked how it would rewrite the ending of one of my favorite TV shows.
But can ChatGPT be used maliciously? ACU Associate professor of computer science Dr. James Prather said the risk is always present.
“It has the potential for great harm if we’re not careful with it,” he said. “The ability to generate large amounts of content that has never existed before— it’s not copying from something—could be harmful in a lot of ways.”
And what about job security?
“I think my students are primarily concerned about ‘Is this going to take my job? Like, if it can write code for me, do I still have a job?’” Dr. Prather said. “I teach computer science. And my response primarily is ‘no.’ Again, it’s just raising the bar. It’s just making you more productive. We didn’t suddenly not need people that can do math when calculators came out. I think this is very similar. Just, on an exponential curve, that now you can be way more productive. You can write more code faster. It enables you to do your job better than you could before.”
The technology is scary. The more information we give it, the more it takes in and learns about us. It’s not quite Rosey from The Jetsons but AI chatbots are here to stay. Distinguishing AI from human text could create challenges in the future, said Chief Product Officer for Turnitin.com Annie Chechitelli.
“I think it is actually difficult for a human to pick up,” she said. “I think that they are looking at a lot of papers very quickly and teachers have a lot of things to do. So, it’s not quite as easy to pick up”
But Dr. Prather believes it’s better we understand this technology now.
“It’s been a big concern here on campus,” he said. “At the beginning of this semester, I think a lot of professors are just scared. Nervous about what this means for their classes. And we held four sessions in our center for teaching and learning here this semester on what AI means for education now in ‘23. And I think by the end of those sessions, people werethey were less scared and more interested in its possibilities.”
