How AI Can Destroy Local Journalism

Image: Shutterstock.com

We all know that local journalism is under extreme pressure. Long established regional newspapers are closing or are being turned into little more than franchise operations where a bare bones local newsroom contributes a modicum of local news to a newspaper fleshed out with filler from national wire services or mother publications. The regional titles in Canada owned by PostMedia are a prime example of this phenomenon. Some digital startups have helped fill the gap, but they too are struggling. There seems to be a reluctance on the part of many to pay for news through subscriptions, while small online publications are forced to compete with everyone from Google and Facebook on down for ad dollars.

One of the supposed remedies, at least in Canada, has been to create various funds to support local journalism. There are a range of programs including the government funded Local Journalism Initiative, launched in 2019 to encourage local news production in “news deserts” and underserved communities, administered by News Media Canada (an industry group), tax credits to offset journalist salaries if the organization is a “Qualified Canadian Journalism Organization” (QCJO), and most recently the Google-funded $100 million fund for media outlets. This was established as a result of Google’s deal with the government to get a five year exemption from being designated under the Online News Act. According to recent reports, the Google fund’s annual contribution to a journalist’s salary will be in the range of C$13,000 to C$20,000 depending on the number of applying journalistic enterprises deemed to qualify.

All these Band-aid measures are designed to staunch the loss of print and digital publications, which has created news deserts in many parts of Canada and the US. In a 2023 report from Northwestern University’s School of Journalism, as reported by Forbes, it was estimated that almost 3000 print newspapers out of approximately 9000 in the US had ceased publication since 2005. The average loss of newspapers in 2023 was 2.5 per week. In my home province of British Columbia alone, Global News reported that the number of daily papers dropped from 36 to just 13 in the six years from 2010 to 2016. According to the Local News Research Project housed at Toronto Metropolitan University’s School of Journalism, over 500 local radio, TV, print and online news operations shuttered in 345 communities across Canada since 2008 although during that time, around 200 new local news outlets, many of them exclusively digital offerings, launched in 152 communities. However, just one opened in 2023. The decline of local news in Canada was well documented in a study released in February of this year by the Public Policy Forum, “The Lost Estate”. The study examines a number of possible remedies, including philanthropic engagement, community foundations, better targetting of government support programs and increased government advertising in local media.

The reasons for the decline of local news are many; the rise of social media, the migration of ad revenues to giant online platforms like Google and META, the reluctance of a younger generation of consumers to pay for news (especially digitally provided content), unauthorized password sharing (even by government!), rising costs of print, and so on. Newly launched digital outlets have tried to fill the gap, but they are facing challenges in getting sufficient ad revenue or subscriptions. In Canada, some will qualify for tax credits or funding from Google, but it will depend on whether they employ at least 2 full time journalists. For many of these digital outlets, their principal modus operandi is to aggregate content from other sources, provide a quick rewrite summary and then insert a link to the original source, thus avoiding copyright infringement. There is some but not a lot of original local journalism.

I have no doubt that with the growing use of AI, some of this initial screening is done through use of artificial intelligence. AI may even be being used to create summaries and rewrites. This saves time and money—but unfortunately cuts down on the need for real journalists. The evaluation of AI’s utility in local journalism is typical of its use in many other areas, from screen writing to auditing to medical diagnoses. It can be a useful tool to enhance productivity, but often at a cost somewhere else, such as with respect to employment. Even if an experienced employee can employ AI to enhance what they are working on, AI could eliminate the beginner or training jobs that help develop the required experience to do this. These challenges are not new and not unique to local journalism.

What is new is the use of AI to take the aggregation model followed by many small local online journals a step further and go nation-wide, a development discovered and highlighted by the Neiman Lab. The Neiman Lab is part of the Neiman Foundation for Journalism, established in 1938 at Harvard. It administers what is proclaimed to be the oldest fellowship program in the world for journalists. The Foundation also publishes a quarterly magazine, Neiman Reports, dedicated to a critical examination of journalism and other journalism-related programs. In an excellent piece of investigative sleuthing, the Lab discovered that “Good Daily”, which operates in 47 states and 355 towns and cities across the US, targeting small town America, is run by just one person, Mathew Henderson, armed with an AI program. If this seems hard to believe, read on.

According to Neiman’s investigation, Henderson operates his “media empire” out of New York. He uses an AI bot to scan the news daily in each local market. The AI program curates the most relevant stories, summarizes them, edits and approves the copy, formats it into a newsletter, and publishes it. The same day! Readers in these 355 towns are led to believe that this is a local publication. It has local testimonials, although the same testimonials, slightly tweaked, appear in various editions around the country. The publications also share the same “About” information and the same mission, which is “to make local news more accessible and highlight extraordinary people in our community.” Henderson claims his automated newsletter is actually helping local publications by driving traffic to them. This is the same argument put forward by META when refusing to pay for news content that it uses on its platform to attract and retain viewers (and sell ads).

Henderson’s business model is to sell advertising and solicit readers for donations. The advertising pie is not infinite, so it is obvious that his “local” newsletters are just one more source of competition for local media chasing ad dollars. Just to be clear, there is nothing illegal about what Henderson is doing. Aggregating content and linking to it is not a violation of copyright law. The lack of full disclosure is a bit disconcerting but I am doubtful if there is anything illegal about having a corporate veil. What is problematic is the cannibalistic nature of his business model, enabled by AI.

This kind of operation, fuelled by AI, can only operate if there is local content to aggregate. The Good Day publications contribute nothing, absolutely zero, to content creation. They, like Facebook, are the ultimate cannibalistic free riders. They can continue to operate successfully, free riding on content created by others, so long as those “others” remain in the business of producing content. However, the more successful businesses like Good Day become, the less viable will be the local journalism sector that produces the content its free-riding competitor subsists on. In the end, the result will be an ouroboros. (Great Scrabble word by the way). The AI driven aggregator will in the end devour the very source of its content. That may be a few years down the road, but logically that is what will happen. It’s like eating your seed grain.

Is there a solution? Many remedies have been proposed, but if there is a silver bullet, I don’t know what it is. For many in the media, finding a way to get the big platforms that benefit from media content to contribute financially to journalism is one avenue, but as we have seen in Canada, Australia, and California, the platforms will pull out all the stops to prevent being required to do so, especially META, which has thumbed its nose at Canada and, now, Australia.

Government subsidies, such as Canada’s Local Journalism Initiative, are resisted by many journalists lest the industry become beholden to government handouts. Holding government to account is one of the key functions of the media, the so-called “Fifth Estate”. Can a subsidized media be trusted to do so? On the other hand, without subsidies will there be any viable local media left? Maintaining its independence is one reason why the small Ottawa-based online publication, Blacklock’s Reporter, that is locked in a David and Goliath struggle with the Government of Canada over the government’s abuse of password-sharing, is opposed to initiatives like the Online News Act. (For the record, that is their view, not mine, but it is a position I respect). Potentially another option could be a tweaking of tax laws to encourage businesses to place ads on local media rather than with the giant international platforms, but in the end business will and should be able to spend its ad dollars where it believes it will get the best results.

For every potential remedy to the problem of keeping local journalism alive, there is a potential downside, just as there is with AI generally. For all its potential advantages AI also has the potential to destroy journalism. Good Daily may be the canary in the coalmine, a fully legal but particularly egregious use of AI putting yet another nail in the coffin of local journalism.

© Hugh Stephens, 2025. All Rights Reserved.