ChatGPT amazes users with its ability to write, but will it change the future of journalism?
By Jordan Scenna
Contributor
When I first read stories out of the tech world promising to eliminate entire sectors of the job market with artificial intelligence, it seemed that truckers would be the first out of work. The evolution of AI would allow automated trucks to drive these highway warriors, these heroes of the long-haul, off our interstates, and park them firmly in the unemployment line. Progress, it seems, is sticking to its word, with many companies either currently using or developing driverless trucks. In fact, a number of industries, from warehousing giant Amazon, to agriculture and finance, are using AI to replace people. Our pesky need to eat and sleep has been a thorn in the side of profit for centuries. Now, the “we do it because we can” attitude of tech wizards is raising the bottom line of businesses everywhere and threatening the livelihood of countless workers.
It never occurred to me that a computer would eventually write me into obsolescence. It’s the factory workers, fast food, and grocery store clerks; they are the casualties of unstoppable technological progress. Besides, writers, like painters, are artists, and machines can’t make art, right? But now, thanks to OpenAI’s ChatGPT software, everyone from journalists to truck drivers are carrying the same load. Will ChatGPT leave writers as antiquated as the quill? Will journalists be usurped by the onslaught of AI?
Can ChatGPT replace journalists?
As someone who yearns to turn writing into a career, when I first experimented with ChatGPT, I was faced with an existential question. “If AI can write a story better than me, faster than me…then what value am I?”
I decided to put ChatGPT to the test. I wanted to see if it could write one of my stories better than I could. I used a story I wrote last semester about a WCC student whose scientific research received recognition. First, I copied my interview notes into ChatGPT. Then, I gave it this prompt: “Write a profile using the following notes in the style of the New Yorker.” It wrote the following lede.
“Luke Thomas is a student at WCC, where he is studying pre-engineering and participating in a science transfer program. He has always been fascinated by astronomy and space science, having grown up as a big fan of the original Star Wars trilogy. Despite his passion for space, his initial college experience at East Tennessee State was marked by poor grades, which led him to drop out.”
Coherent enough, but pretty bland. No flavor, no sauce, nothing much to capture the reader’s attention. So, I tasked it with spicing things up. My second ask was to “write the same article with an interesting anecdotal lede.” Seconds later, ChatGPT came back with,
“Luke Thomas is the epitome of what can happen when someone sets their sights on a goal and refuses to give up. From a college dropout with bad grades, to an aspiring astrophysicist with a published paper to his name, Thomas’ journey is a testament to the power of determination and hard work.”
This was a marked improvement. By just tweaking what I asked ChatGPT to do and building on its previous response, it wrote a better lede and it did it in seconds. Not yet worthy of the New Yorker, but a recognizable leap.
For my last experiment, I wanted to see if it could make me a nationally published writer. I input my entire article and said, “Improve this to be publishable in Vanity Fair.” When ChatGPT cut my article in half, I couldn’t help but feel it was calling me a windbag without actually calling me a windbag; the thing even has decorum. But to my surprise, and great relief, the resulting profile was a worse version than its previous attempts.
In its current form, ChatGPT can’t replace journalists, not fully anyway (with a little who, what, when, and where, it can write you a hard news story without so much as a sip of coffee). For one thing, it can’t run an interview. It can’t read a person, pick up on cues, and know what to ask and what not to. It also has a lying problem. When I asked ChatGPT to write a bio on me, for some inexplicable reason, it didn’t know who I was, so it just made something up—although a glowing bio it was–apparently I’m a professional soccer player who played his college ball at Michigan State University. It also can’t understand context and the subtlety of human emotion and experience, which is indispensable to great storytelling.
Today, for budding journalists, ChatGPT is the monster lurking under the bed. It can’t hurt us yet, but with society’s fear of technological stagnation, it’s getting stronger. OpenAI CEO Sam Altman has already said that a fourth-generation model of ChatGPT may be released this year. Altman proposed that a new model may have the capability to learn and improve on its own, something that GPT-3.5 can’t do. This means once GPT-4 is online, the model can self-upgrade and improve based on its interactions, becoming a better version of itself.
AI language models are still in their infancy, but have made considerable progress in the last few years. It’s only a matter of time before they come for, not just journalism jobs, but many jobs in art related fields. One day, your computer will be able to sit you down for an interview. It will call you, or you’ll be able to login (whenever you want because computers don’t sleep, or eat, or get tired) and it’ll ask you the tough questions without reservation or bias.
It’s a brave new world, and human resources are going digital. Just not quite yet.
But don’t take my word for it, just ask ChatGPT.
-Can ChatGPT replace journalists?
“…Journalism requires human judgment, critical thinking, and a deep understanding of the world and the people who live in it.”
Well said.
