Tag: artificial intelligence

The triumph of hope over experience: The latest on how AI is not solving the local news crisis

Illustration produced by AI using DALL-E

This past weekend I listened to a bracingly entertaining conversation that the public radio program “On the Media” conducted with tech journalist Ed Zitron. Co-host Brooke Gladstone had billed it as a chance for Zitron to make sense out of DeepSeek, the new Chinese artificial-intelligence software that purports to do what ChatGPT and its ilk can do for a fraction of the cost — and, presumably, while using a fraction of the electric power burned by American AI companies.

But it was so much more than that. Maybe you’re familiar with Zitron. I wasn’t. As I learned, he is a caustic skeptic of American AI in general. In fact, he doesn’t even regard the large language models (LLMs) that we’ve come to think of as AI as the real thing, saying they are nothing but an error-prone scam that is attracting fast sums of venture capital but will never make any money. Here’s a taste:

The real damage that DeepSeek’s done is they’ve proven that America doesn’t really want to innovate. America doesn’t compete. There is no AI arms race. There is no real killer app to any of this. ChatGPT has 200 million weekly users. People say that’s a sign of something. Yes, that’s what happens when literally every news outlet, all the time, for two years, has been saying that ChatGPT is the biggest thing without sitting down and saying, “What does this bloody thing do and why does it matter?” “Oh, great. It helps me cheat at my college papers.”

And this:

When you actually look at the products, like OpenAI’s operator, they suck. They’re crap. They don’t work. Even now the media is still like, “Well, theoretically this could work.” They can’t. Large language models are not built for distinct tasks. They don’t do things. They are language models. If you are going to make an agent work, you have to find rules for effectively the real world, which AI has proven itself. I mean real AI, not generative AI that isn’t even autonomous is quite difficult.

As you can tell, Zitron has a Brit’s gift for vitriol, which made the program all the more compelling. Now, I am absolutely no expert in AI, but I was intrigued by Zitron’s assertion that LLMs are not AI, and that real AI is already working well in things like autonomous cars. (Really?) But given that we just can’t keep AI — excuse me, LLMs — from infesting journalism, I regarded Gladstone’s interview with Zitron as a reason to be hopeful. Maybe the robots aren’t going to take over after all.

Continue reading “The triumph of hope over experience: The latest on how AI is not solving the local news crisis”

A flick of the mutant wrist

Adam Gaffin has posted a hilarious find at Universal Hub — an AI-generated X-ray of a wrist published by The Boston Globe that has all kinds of problems, including a third forearm bone and finger bones that don’t actually connect to anything. It’s now been online since Dec. 19, and it persists despite some mockery on Bluesky as well as Adam’s post.

Playing with AI: Can Otter and ChatGPT produce a good-enough account of a podcast interview?

This post will no doubt have limited appeal, but a few readers might find it interesting. I’ve been thinking about how to produce summaries and news stories based on the podcast that Ellen Clegg and I host, “What Works: The Future of Local News.” The best way would be to pay a student to write it up. But is it also a task that could be turned over to AI?

Purely as an experiment, I took our most recent podcast — an interview with Scott Brodbeck, founder and CEO of Local News Now, in the Virginia suburbs of Washington, D.C. — and turned it over to the robots.

I started by downloading the audio and feeding it into Otter, a web-based transcription service that uses AI to guess at what the speaker might actually be saying. Once I had a transcript, I took a part of it — our conversation with Brodbeck, eliminating the introduction and other features — and fed it into ChatGPT twice, once asking it to produce a 600-word summary and then again to produce a 600-word news story. Important caveat: I did very little to clean up the transcript and did not edit what ChatGPT spit out.

The results were pretty good. I’m guessing it would have been better if I had been using a paid version of ChatGPT, but that would require, you know, money. I’d say that what AI produced would be publishable if some human-powered editing were employed to fix it up. Anyway, here are the results.

The transcript

Q: Scott, so many of the projects that we have looked at are nonprofit, and that trend seems to be accelerating. In fact, we love nonprofit news, but we also worry that there are limits to how much community journalism can be supported by philanthropy. So your project is for profit. How have you made that work? Dan, do you think for profit? Digital only, local news can thrive in other parts of the country as well. Continue reading “Playing with AI: Can Otter and ChatGPT produce a good-enough account of a podcast interview?”

Northeastern news project wins $100k grant; plus, more on the Herald, and AI hell in Melrose

We have some exciting news about one of our sister projects at Northeastern University’s School of Journalism. The Scope, a professionally edited digital publication that covers “stories of hope, justice and resilience” in Greater Boston, has received a $100,000 grant from Press Forward, a major philanthropic initiative funding local news.

“Since its launch in late 2017, The Scope has become a national leader in leveraging university resources to help solve the news desert crisis. This grant is a vote of confidence in our model,” said Professor Meg Heckman in the announcement of the grant. “Rebuilding the local information ecosystem is a big job, and we’re thrilled Press Forward sees the School of Journalism as a vital part of the solution.”

Heckman has been the guiding force behind The Scope for several years now. Joining her in putting the grant application together were the school’s director, Professor Jonathan Kaufman, and Professor Matt Carroll.

The Scope was one of 205 local news outlets that will receive $20 million in grant money, according to an announcement by Press Forward on Wednesday. Several of the projects are connected in one way or another to What Works, our project on the future of local news:

• Santa Cruz Local (California), which competes with a larger and better-known startup called Lookout Santa Cruz. Santa Cruz Local co-founder Kara Meyberg Guzman and Lookout Santa Cruz founder Ken Doctor were both interviewed for the book that Ellen Clegg and I wrote, “What Works in Community News,” as well as on our podcast, “What Works: The Future of Local News.”

• The Boston Institute for Nonprofit News, an investigative project that publishes stories on its own website as well as in other outlets. Co-founder Jason Pramas has been a guest on our podcast. Several other Boston-based outlets received grants as well: the Dorchester Reporter, a 40-year-old weekly newspaper; Boston Korea, which serves the Korean American Community in Massachusetts, Rhode Island and New Hampshire; and El Planeta, a venerable Spanish-language newspaper.

• The Maine Monitor, a digital project that covers public policy and politics. Now-retired editor David Dahl has been a guest on our podcast.

• InDepthNH, published by the New Hampshire Center for Public Interest Journalism. The site focuses on public policy and politics, and its founder, Nancy West, has been a podcast guest.

• Montclair Local (New Jersey), a hyperlocal website that is one of the projects we write about in “What Works in Community News.” In 2009, the Local merged with Baristanet, one of the original hyperlocal news startups, which I wrote about in my 2013 book, “The Wired City.”

• Eugene Weekly (Oregon), an alternative weekly that suffered a near-death experience earlier this year after a former employee embezzled tens of thousands of dollars. I wrote about that here and at our What Works website.

More on the shrinking Herald

Earlier this week I wrote about the latest paid circulation figures for the Boston Herald based on its recent filings with the U.S. Postal Service. I lamented that the numbers weren’t as complete as I would have liked because the Alliance for Audited Media was no longer providing its reports for free to journalists and researchers, as it had done in the past.

Well, it turns out that I was knocking on the wrong door. I now have recent reports for both the Herald and The Boston Globe. The AAM figures don’t significantly change what I reported about the Globe, but they do fill in some gaps for the Herald.

For March 2024, the most recent AAM report that’s available, the Herald’s average weekday paid print circulation for the previous six months was 12,272, a decline of 2,247, or nearly 15.5%, compared to its March 2023 totals. Sunday paid print circulation, according to the March 2024 report, was 15,183, down 2,690, also 15%.

As I explained earlier, AAM tallies up paid digital circulation differently from a newspaper’s internal count; among other things, AAM allows for some double-counting between print and digital. Nevertheless, its digital figures are useful for tracking trends.

In the March 2024 report, according to AAM, the Herald’s total average weekday paid digital circulation was 30,009, which actually amounts to a decrease of 2,250, or about 7%, over the previous year. Sunday paid digital in March 2024 was 29,753, down 1,952, or about 6.1%.

Needless to say, that’s not the direction that Herald executives want to be moving in — although I should note that, in its September 2024 post office filing, the Herald reported a slight rise in its seven-day digital circulation compared to the previous year.

What fresh hell is this?

The Boston suburb of Melrose is not a news desert. It has a newspaper, the Melrose Weekly News. But, like many communities, it would benefit from more news than it’s getting now, especially after Gannett shuttered the venerable Melrose Free Press in 2021.

So … artificial intelligence to the rescue? In CommonWealth Beacon, Jennifer Smith introduces us to the “Melrose Update Robocast,” which uses fake voices, male and female, to talk about local issues based on information that’s fed into it to produce an AI-generated script. (Note: Smith interviewed me for the piece, though I didn’t make the cut. I’m also on CommonWealth’s editorial advisory board.)

“In a way, what I’m talking about is an act of desperation,” “Robocast” creator Tom Catalini tells Smith.

Yet all across Massachusetts, independently operated news sites with real human beings are springing up to cover local news. Community journalism is how we connect with each other, and an AI-generated podcast can’t do that.

In Medford, where I live, we haven’t had a local news source for two years. But we do have a podcast, “Medford Bytes,” hosted by two activist residents who convene important conversations about what’s going on in the city, including a recent interview with the mayor about three contentious ballot questions that would raise taxes in order to pay for schools, road repairs and a new fire station.

That’s the sound of community members talking among themselves.

A conversation with ChatGPT, given a voice by Siri, shows AI’s promise and limitations

Mike Blinder, publisher of the trade journal Editor & Publisher and the host of its vodcast, “E&P Reports,” tried something unusual recently. He hooked up ChatGPT to Siri and conducted an interview about issues related to artificial intelligence and journalism. The result is like a smarter version of Eliza, a 1960s-vintage AI program that could carry on what seemed like a realistic conversation.

Blinder has become something of an AI evangelist, using to automate some of E&P’s editorial processes and asking ChatGPT to write bios of guests on “E&P Reports.” But AI has a long way to go in terms of carrying on an intelligent conversation that’s also spontaneous.

For instance, as you’ll hear, Blinder’s approach in interviewing ChatGPT is to lead the witness with long, information-packed questions that the chatbot can then use to scour the internet and come back with a plausible-sounding answer.

I also detected at least one error. In response to a question about the possibility that AI-powered search engines will harm news organizations by removing any incentive to click through, ChatGPT says:

When AI tools like Perplexity or Chat GPT provide detailed summaries of content without directing traffic back to the original source, it creates a real challenge for news publishers who rely on page views for ad revenue and subscriptions. I completely understand the concern.

The problem is that Perplexity actually does cite its sources, which differentiates it from ChatGPT and other competitors. It’s why I suggest to my students that Perplexity is a useful tool as long as they click through, and it’s why I use it as well.

Nevertheless, Blinder’s close encounter of the robotic kind is fun and interesting. You can watch it on YouTube or subscribe on any podcast app. Blinder’s been a guest on our “What Works” podcast about local news twice, and Ellen Clegg and I were on “E&P Reports” earlier this year to talk about our book, “What Works in Community News.”

Gannett to lay off 74 employees in Mass. as it prepares to shut down its consumer site

Gannett and USA Today headquarters in McLean, Va. Photo (cc) 2008 by Patrickneil.

Gannett is laying off 74 employees in Massachusetts — but, for once, they are not people who were producing local journalism. The layoffs, which take effect Nov. 14, are related to the company’s decision to close Cambridge-based Reviewed, a website that combines consumer advice and commerce in a manner similar to Wirecutter, which is part of The New York Times.

The pending closure and layoffs were reported Aug. 26 by Mia Sato at The Verge and came amid accusations that Reviewed published articles produced by artificial intelligence and attributed to non-existent writers. Sato wrote: “As The Verge reported last fall, the marketing firm behind the Reviewed content is the same company that was responsible for a similar dust-up at Sports Illustrated, in which remarkably similar product reviews were published and attributed to freelancers.”

Gannett denied the allegations and said the decision to shut down Reviewed was based on changes in Google’s algorithms.

Aidan Ryan of The Boston Globe quotes NewsGuild of New York president Susan DeCarava in a statement:

We are deeply troubled by Gannett’s decision to shutter Reviewed. We are concerned for the future of dozens of workers represented by The NewsGuild of New York working at Reviewed, and about the broader impact of this announcement on the media industry at large.

The layoffs were announced in advance, reports Ray Schultz of Publishers Daily, because of a Massachusetts law mandating that companies provide 60 days’ notice ahead of a mass layoff.

Earlier:

 

AI translation is a boon for the Vineyard’s newspapers, but there are limitations as well

Martha’s Vineyard. Photo (cc) 2012 by David Berkowitz.

The two independent weekly newspapers that cover Martha’s Vineyard are using AI-powered translation software to provide the island’s growing Brazilian population with Portuguese-language articles, Aidan Ryan reports in The Boston Globe.

But though the service is surely a step forward, stronger coverage of the Brazilian community still depends on the human touch, MV Times publisher Charles Sennott and Vineyard Gazette publisher Monica Brady-Meyerov acknowledge. Ryan writes:

While these initiatives can reflect a sincere effort to serve the public with reliable information, they also often highlight the limitations of technology and gaps in staff diversity for many traditional English-language newsrooms.

Reliable automated translation can be a boon. The English-language New Haven Independent, for example, relies on people power to provide one translated article each week to run in La Voz Hispana de Connecticut, a weekly Spanish-language paper with which it shares a newsroom. A translated cartoon from La Voz runs in the Independent’s newsletter. (Both projects are featured in our book, “What Works in Community News.”)

AI could provide readers of both publications with much more than they’re getting now.

Wyoming reporter resigns after confronted with evidence that he used AI to pipe quotes

Buffalo Bill (in wide-brimmed hat), the purported founder of the Cody Enterprise, in 1912 or thereabouts. Photo uploaded by Alden Jewell.

A reporter in Wyoming has resigned after he was caught using artificial intelligence to make up quotes, with AI pulling in material that had been published years earlier and producing sentences such as “This structure ensures that the most critical information is presented first, making it easier for readers to grasp the main points quickly” — an apparent description of the inverted-pyramid form of newswriting.

Aaron Pelczar, a reporter for the Cody Enterprise (“Founded by Buffalo Bill in 1899”), quit after CJ Baker of the rival Powell Tribune confronted him with the goods.

“To date,” Baker wrote, “seven people — ranging from Gov. Mark Gordon to the victim of an alleged crime — have indicated to the Tribune that they didn’t make the statements Pelczar quoted them as making. The Tribune also found a number of other quotes that were altered in some way or attributed to the wrong person.”

Baker’s account is just nuts and is well worth spending a few minutes poring over. And thanks to Adam Gaffin of Universal Hub for passing this along.

Newsletters move to the fore as tech platforms spurn community journalism

1923 photo via the Library of Congress

If we’ve learned anything about news publishing in recent years, it’s that the giant tech platforms are not our friends. Google is embracing artificial intelligence, which means that searching for something will soon provide you with robot-generated answers (right or wrong!), thus reducing the need to click through. Facebook is moving away from news. Twitter/X has deteriorated badly under the chaotic leadership of Elon Musk, although it still has enough clout that President Biden used it to announce he was ending his re-election campaign.

So what should publishers do instead? It’s no secret — they’re already doing it. They are using email newsletters to drive their audience to their journalism. A recent post by Andrew Rockway and Dylan Sanchez for LION (Local Independent Online News) Publishers reports that 95% of member publishers are offering newsletters, up from 81% in 2022. “The decline in referral traffic,” they write, “will likely lead to more direct engagement by publishers with their audiences.”

Some observers worry about newsletter overload as our inboxes fill up with email we may never get around to reading. That’s potentially a problem, but I think it’s a more serious problem for larger outlets, many of which send out multiple newsletters throughout the day and risk reaching a point of diminishing returns. By contrast, users will value one daily newsletter from their hyperlocal news project with links to the latest stories.

Newsletters are crucial to the success that Ellen Clegg and I have seen both in the projects we write about in our book, “What Works in Community News,” and on our podcast, “What Works: The Future of Local News.” Essentially, we’ve seen three newsletter strategies.

  • By far the most common approach publishers use is to offer a free newsletter aimed at driving users to their website, which may be free or subscription-based. The Massachusetts-based Bedford Citizen, for instance, sends out a daily newsletter generated by its RSS feed and a weekly human-curated newsletter. The Citizen is a free nonprofit, but once they’ve enticed you with their top-of-the-funnel newsletter, they hope they can lure you into becoming a paying member. Ellen and I interviewed executive director Teri Morrow and editor Wayne Braverman on our podcast last February.
  • The Colorado Sun, a statewide nonprofit, offers a series of free and paid newsletters, while the website itself is free. The paid newsletters represent an unusual twist: Some of them feature deeper reporting than you can get from the website on topics such as politics, climate change and outdoor recreation. At $22 a month for a premium membership, users pay no more than they would for a digital subscription to a  daily newspaper. Editor Larry Ryckman talked about that in our most recent podcast.
  • In some places, the newsletter is the publication. An example of that is Burlington Buzz, a daily newsletter that covers Burlington, Massachusetts. Founder, publisher and editor Nicci Kadilak recently switched her newsletter platform from Substack to Indiegraf, and her homepage looks a lot like a standard community website — which shows that it’s a mistake to get too caught up on categories when newsletters have websites and websites have newsletters. Ellen and I talked with Nicci last year.

What’s crucial is that news publishers have direct control of the tools that they use to connect with their audience. Gone are the days when we could rely on Facebook and Twitter to reliably deliver readers to us. We have to go find them — and give them a reason to keep coming back.

Correction: Burlington Buzz has moved to Indiegraf, not Ghost.

Leave a comment | Read comments

Reuters exposes NewsBreak’s Chinese ties and AI-based local news hallucinations

Image (cc) 2019 by Journolink

You would have thought that misusing artificial intelligence to make up a murder that never took place and to amplify a false story about yet another murder would have been enough to stop NewsBreak. But the media outlet, which has rushed in to fill the gap caused by the decline of legitimate local news, has continued along its hallucinatory way — until now.

Last week, Reuters published an investigative report into NewsBreak exposing the company’s Chinese ties and its use of AI to write fiction. Reporter James Pearson begins with the fake murder last Christmas Eve from Bridgeton, New Jersey, which I wrote about in January. Here’s Pearson’s lead:

Last Christmas Eve, NewsBreak, opens new tab, a free app with roots in China that is the most downloaded news app in the United States, published an alarming piece about a small town shooting. It was headlined “Christmas Day Tragedy Strikes Bridgeton, New Jersey Amid Rising Gun Violence in Small Towns.”

The problem was, no such shooting took place. The Bridgeton, New Jersey police department posted a statement on Facebook on December 27 dismissing the article — produced using AI technology — as “entirely false.”

“Nothing even similar to this story occurred on or around Christmas, or even in recent memory for the area they described,” the post said. “It seems this ‘news’ outlet’s AI writes fiction they have no problem publishing to readers.”

Pearson fails to mention that the fake story was first exposed by Eric Conklin of NJ.com. Nor does he report that, just a few weeks earlier, NewsBreak used AI to amplify yet another fake-murder story that was produced by actual humans at the Mid Hudson News in New York, a monumental error revealed by Lana Bellamy and Phillip Pantuso of the Times Union, based in Albany.

Despite the lack of credit to local news outlets, the Reuters investigation represents the deepest dive yet into NewsBreak, and could result in action. As Pearson notes, NewsBreak, despite its sleazy tactics, has content-sharing arrangements with news outlets such as The Associated Press, CNN, Fox and, well, Reuters. As far back as 2022, Norman Pearlstine, a longtime top news executive who was working as a consultant for NewsBreak, warned his client: “I cannot think of a faster way to destroy the NewsBreak brand.”

NewsBreak remains a popular destination for people looking for local news. According to SimilarWeb, it was visited nearly 57 million times between March 24 and May 24. But maybe NewsBreak’s reign of error is coming to an end. In a follow-up story, members of Congress voiced concern, with Sen. Mark Warner, chair of the Intelligence Committee, saying, “The only thing more terrifying than a company that deals in unchecked, artificially generated news, is one with deep ties to an adversarial foreign government.”

To be fair, the company says those ties were severed several years ago. Still, NewsBreak’s irresponsibly sloppy use of AI to generate fake local news needs to be called out and shamed.

Leave a comment | Read comments