How AI is changing the journalism industry

In the age of Hollywood writers striking over the use of artificial intelligence in writers’ rooms and AI-written articles, many have expressed concerns over the capabilities of artificial intelligence to replace journalists and media producers alike. Instead, some see artificial intelligence as a tool to expand the capabilities of the creative and journalistic industries. 

For David Trilling, co-founder of the AI-generated Inside Arlington newsletter, large language models have the potential to fulfill the growing lack of news coverage in local communities. 

“In a town of Arlington’s size [~46,000] … there was still very little local media about what was going on in town … we do have a wonderful local online paper that does its best, but it’s severely underfunded and understaffed,” said Trilling. 

Trilling began the newsletter with co-founder Winston Chen after Chen wanted to make changes in leash laws in the community, but did not have the time or resources to be informed about community meetings. Inside Arlington uses a combination of preexisting artificial intelligence tools, such as ChatGPT, which are then trained on recorded meetings and meeting minutes to create brief newsletters detailing the events that occurred at community meetings. While the newsroom relies on AI to write the articles, the technology does not replace journalists, according to Trilling.

“While the newsroom relies on AI to write the articles, the technology does not replace journalists, according to Trilling.”

 “I don’t think this is replacing journalists because at least in our case, there were no journalists doing this … no one is covering these meetings,” said Trilling.

Others are looking to AI to assist in the editing process of local newsrooms. Michelle Johnson, a fellow at Boston University’s Hub for Civic Tech Impact, is currently developing Editor AI, a tool that aims to help local journalists edit their work to the standards of their newsrooms. The product is still in development, but Johnson hopes to target small, local publications, who may otherwise be limited in their capabilities.

 “This is not for big audiences, it’s not for the Boston Globe. They have money and people to build custom stuff, proprietary stuff. I’m talking about some little publication in a small town where there’s one person,” she said.

As AI is being deployed to help writers and editors, the possibility of an AI-takeover of creative industries is not far off. While the technology is not yet able to replicate the work done by journalists, it serves as a cheaper and quicker solution in an increasingly shrinking news industry.

Some newsrooms have even begun to deploy autonomously written articles. Media giant Gannett recently came under fire after the company was forced to pause its use of Lede AI, a sports writing AI tool, after the technology made multiple errors in stories about high school athletics. However, this issue will continue to persist as long as companies value profits over products, according to communications expert Julie Rafferty. 

“The world is driven predominantly by money, and not by accuracy and creativity. If a company [can] feel like [it] can get something that’s good enough using AI rather than AI plus humans, jobs will be lost, quality will decline,” said Rafferty, founder of Rafferty Communications and a former journalist.

Though artificially intelligent large-language models have strong capabilities, they are limited in their nuance. Essentially, the large-language models behind AI chatbots like ChatGPT are trained to identify patterns in the mechanics of writing through viewing a variety of writing samples gathered from a broad array of sources, including news articles. This gives these models limited ability to create factual stories without human oversight.

“A language model is … like a giant autocomplete system that’s finding statistical patterns in language and trying to predict the next word over and over again until it has something coherent,” said Jack Bandy, professor of computer science at Transylvania University in Lexington, Kentucky. 

Ultimately, the future of artificial intelligence in newsrooms is up to journalists, according to Bandy.

“ ‘There are implications and there are applications of new technology, but people get to decide what those are. It’s not guaranteed to be a doomsday for journalists or to put all journalists out of jobs … journalists will get to decide how they use it and what they want to do with it,’ said Jack Bandy.”

 “There are implications and there are applications of new technology, but people get to decide what those are. It’s not guaranteed to be a doomsday for journalists or to put all journalists out of jobs … journalists will get to decide how they use it and what they want to do with it,” he said.