Using ChatGPT to come up with a weird news story, and then interviewing the Kenosha Steampunk Society – TMJ4 News

January 19, 2024 ndowd

Menu
KENOSHA — Instead of coming up with a story idea by myself, I asked ChatGPT to come up with one for me. I asked ChatGPT for ‘weird story ideas in Milwaukee’. The AI language program gave me a lot of ideas, but one stood out among the rest.
ChatGPT is an AI tool that provides detailed and human-like responses to questions.
In this case, it responded with, “The quirky and unusual subcultures of Milwaukee such as the roller derby community or the steampunk community.” TMJ4 has done stories on roller derby communities before, but never on the steampunk community. After a quick Google search, I found the Kenosha Steampunk Society, got in my car, and headed south. That’s where I met Paul Little.
“Well a lot of people know me as that steampunk guy,” Little said.
He leads the Kenosha Steampunk Society, and he owns a pop-up that sells steampunk gear. He describes steampunk as, “It’s a very creative revisionist retro-futurism.”
In other words, it’s a fantasy Victorian-era aesthetic where people dress as what those who lived during the Industrial Revolution in the 1800s thought people would dress like in the future. Steampunk incorporates lots of gears and steam-powered imagery into its aesthetic.
“You use a lot of old stuff like recycling. So you’re not throwing away stuff. It’s really big into history,” Little said.
Many people thrift their costumes or re-use old fabrics to create their designs.
“Seeing people’s creativity like honestly it’s like a walking fashion show the stuff that people come up with,” Danielle Lieber said.
Lieber is an author who writes fantasy, steampunk, and romance novels. She is also involved in the Kenosha steampunk community. Lieber is the co-lead of the local Tea Scout troop. It’s a group of people that come together to drink tea, share steampunk culture, and raise money for local organizations.
“We might look a little weird, but weird makes the world go round I think,” Lieber said.
The Kenosha Steampunk Society meets regularly at pop-ups hosted by Little or at the Tea Scout gatherings. There are other steampunk communities in Milwaukee, Janesville, Baraboo, Madison, and Chicago.
“I love the fantasy of it all I love the inclusiveness I love how they really get people to come out of their shells,” Nicholas Cowart, who was dressed as a steampunk clown and runs an entertainment company, said.
The steampunk community is definitely one that is not covered or mentioned often in the Wisconsin news cycle, so how did ChatGPT know about them and decide it was a good suggestion?
“It doesn’t really. I mean what it does know is word associations because it has ingested boatloads – we’re talking millions – of words of text,” Susan McRoy, the chair of the Computer Science Department at UW-Milwaukee, said.
That means that ChatGPT didn’t really know what would make a good story. It just made an educated guess based on internet data. Most of the story ideas ChatGPT gave me were either not real, to broad, or happening in places like Door County.
In fact, I had to do many Google and Facebook searches before I found the Kenosha Steampunk Society. It’s a small group that doesn’t have a robust digital footprint.
“You did the real work, but it gave you an idea, and that’s useful as long as you don’t use it to attribute facts,” McRoy said.
The answers ChatGPT don’t have to be accurate. The point isn’t that the program will provide you with 100% true statements about anything you ask it. The goal is that it can answer your question in highly plausible ways. ChatGPT combs the internet for corresponding words that seem correct and generates a response.
“That combination of a description, ‘oh an unusual experience’ and ‘steampunk’ or whatever has probably occurred more often than some other combination,” McRoy said.
That’s how ChatGPT gave me that answer.
Overall, this was a fun experiment. There absolutely are ways that AI and local news can collaborate, but kinks still need to be worked out before the two can truly mesh. That’s probably the case for many industries.