As local newspapers grapple with shrinking budgets and overburdened journalists, some newsrooms are experimenting with an idea that skeptics say threatens the very role of reporters: integrating artificial intelligence into the newsroom.
Editors remain cautious about the use of AI in reporting, one of the main reasons for which is that it cannot distinguish fact from fiction. But used responsibly, they say, it can provide a cost-effective set of tools to lighten the burden on local journalists and expand their reach, such as through AI-generated city council meeting summaries.
Why did we write this
As local news organizations shrink or disappear, journalists are turning to artificial intelligence to fill the gap. Can you trust an AI that can’t tell truth from fiction?
Renee Richardson, Editor-in-Chief of Brainerd Dispatch in Brainerd, Minnesota, is an AI integration pioneer in her local newsroom. Their artificial intelligence experiment will begin in June to automate public safety announcements from police bulletins. Ms. Richardson hopes to maximize the efficiency of her staff’s workflow and give the dispatch reporters back something priceless: time.
“We are constantly asking our employees to do more and provide more information in a variety of ways. Whether it’s social media, video podcasts, audio segments, all of our photos, or all the parts that go into them. We rarely do anything that turns back time for them. The advantage I see in this is to finally give them back that time.”
As local newspapers grapple with shrinking budgets and overburdened journalists, some newsrooms are experimenting with an idea that skeptics say threatens the very role of reporters: integrating artificial intelligence into the newsroom.
Editors remain cautious about the use of AI in reporting, one of the main reasons for which is that it cannot distinguish fact from fiction. But used responsibly, they say, it can provide a cost-effective set of tools to lighten the burden on local journalists and expand their reach, such as through AI-generated city council meeting summaries.
Silicon Valley AI company OpenAI helped spark interest and debate around writing and reporting on AI by releasing its conversational chatbot ChatGPT in late 2022. The AI-based program can quickly respond to text commands, and then write essays, summarize books. and prepare financial reports. Its release attracted nationwide attention and additional funding from Microsoft.
Why did we write this
As local news organizations shrink or disappear, journalists are turning to artificial intelligence to fill the gap. Can you trust an AI that can’t tell truth from fiction?
In California’s Humboldt County, 300 miles north of Silicon Valley, Hank Sims and his local Lost Coast Outpost newsroom began experimenting with ChatGPT last year. The web-only newsroom used the program to develop its own version, dubbed LoCOBot. The program downloads and summarizes the agendas of local community meetings.
Mr. Sims says LoCOBot replaces the human need to review planned lengthy agendas for city council and other meetings, and frees up journalists to investigate larger events.
“Our reporters love this because it automatically notifies them when a new agenda is published and gives a kind of sneak peek from our side of what the board will be talking about,” says Mr. Sims.
LoCOBot is open to the public and easily accessible through the Lost Coast Outpost. website. In addition to summarizing agenda items in concise professional language, LoCOBot can turn dry bureaucratic prose into humor and poetry.
For example, there was a Cinderella moment on the agenda of a city council meeting in the small town of Arcata, when LoCOBot turned it into a bedtime story: “A long time ago, there was a CFO named Tabata in the city of Arcata. She was a hardworking and dedicated employee in charge of overseeing the city’s budget.”
In all likelihood, the CFO is indeed “hardworking and dedicated.” But of course, LoCOBot has never interviewed the director or looked into her track record, making this fun fairytale presentation something of a cautionary tale about AI’s shortcomings. Many experts say that real journalists who ask questions and follow the news still play an important role in impartial reporting of the news and uncovering the truth.
Conversational chatbots or large language models like ChatGPT can’t distinguish between true and false, explains Nir Eisikowitz, director of the Center for Applied Ethics at the University of Massachusetts Boston. He says they can’t be trusted because there are no built-in guardrails. To use AI responsibly in newsrooms, you always need a reporter who double-checks AI-generated work.
“I think it’s ironic that journalism, a profession whose raison d’etre is to uncover the truth, relies on a tool that can’t tell truth from lies,” says Mr. Eisikowitz. “The more important the story and the higher the stakes, the less I will rely on AI.”
According to Mr. Sims, despite such concerns, Lost Shore Outpost had no problem with LoCOBot misinformation or accuracy.
This is mainly due to the fact that LoCOBot has simple direct commands for summarizing texts rather than creating open responses. He adds that their publication is prudent and intentional, with author’s signatures transparently identifying the work done by the program.
As part of an ongoing effort to help local newsrooms integrate automation and AI technology, the Associated Press and other news organizations are funding the spread of AI technology across the country, from Michigan to Puerto Rico.
In Michigan, reporter Dustin Dwyer is responsible for coverage of nearly the entire western Great Lakes state for Michigan Radio’s WUOM-FM.
When newspapers flourished, Mr. Dwyer relied on local news as a resource. But as the number of newspapers dwindled, Mr. Dwyer was exhausted and his coverage became strained, he says.
“For example, in Cedar Springs, Michigan, their little newspaper just closed,” Mr. Dwyer explains. “Right now, no one is covering Cedar Springs city council meetings, and I can’t go there. So is there another way to get coverage for all those other localities? Is there a way to still give people the tools they need to inform their encounters with the realities we face in local news and recruiting?”
For the past three years, Mr. Dwyer and his fellow reporters at Michigan Radio have used their own developed program called Minutes to record city council and subcommittee meetings in more than 40 cities across the state. Minutes transcribed over 5,000 recordings and provided transcripts to reporters.
Mr. Dwyer says AI accuracy remains a major challenge. Transcriptions help reporters look up key words and phrases, but the protocols are far from reliable, meaning that fast-talking and grouchy phrases are prone to transcription errors. Yet Mr. Dwyer sees AI as a viable tool for journalists, not a replacement. “It allows a reporter like me to reach a much wider audience than I could do it myself.”
With the help of the AP, Michigan Radio will enhance and enhance the transcription capabilities of Minutes audio with OpenAI’s Whisper speech-to-text program.
Renee Richardson, Managing Editor of Brainerd Dispatch in Brainerd, Minnesota, is the latest AI integration pioneer in her local newsroom. Their artificial intelligence experiment will begin in June.
By working with a developer to create an artificial intelligence program to automate public safety announcements from police bulletins, Ms. Richardson hopes to maximize the efficiency of her staff’s workflow and give reporters back something priceless: time.
“We are constantly asking our employees to do more and provide more information in a variety of ways. Whether it’s social media, video podcasts, audio segments, all of our photos, or all the parts that go into them. We rarely do anything that turns back time for them. The advantage I see in this is to finally give them back that time.”
And the idea that AI might one day make reporters extinct?
“When do you think: “Is this something that will replace my job?” the answer for us is: no, it will just free you up for even better and more efficient work.”