The Great AI News Heist
The latest turn in Big Tech's theft of journalists' labor.
This is Boondoggle, the newsletter about how corporations rip off our states, cities, and communities, and what we can do about it. If you’re not currently a subscriber, please click the green button below to sign up. Thanks!
A group of 400 local newspaper publishers recently sued OpenAI and Microsoft, alleging that their artificial intelligence-powered chat programs are the product of (among many other things) widespread, rampant theft from the news industry.
In order to train their flagship products — ChatGPT and Copilot, respectively — to ably answer users’ queries, OpenAI and Microsoft broadly scraped the internet and used all the content they grabbed as training material, without paying any of the people who created that content. Of course, a meaningful percentage of that content was news articles produced by working journalists, which was then baked into the answers AI programs provide to users, with most of the attribution and, more importantly, all of the payments stripped out, in violation of U.S. copyright law.
As the complaint notes, “By covertly scraping, copying, adapting, and using the Publishers’ content to train and deploy their GenAI products, Defendants have diverted traffic from the Publishers’ news sites, thereby causing a reduction in the Publishers’ advertising and subscription revenue while denying the Publishers the licensing fees the Defendants should have been required to pay to use the Publishers’ protected works.”
Indeed, ChatGPT and programs like it only function because of mass theft from journalists (and authors, and musicians, etc.). OpenAI’s founder testified before the British House of Lords that it is “impossible to train today’s leading AI models without using copyrighted materials,” so that’s exactly what they did, only later making some limited payments to copyright holders that haven’t made most journalists (or authors or musicians or whoever else) whole.
The suit, led by former New Jersey Attorney General Matt Platkin, covers small news publishers across the country, from New York to Iowa to Arkansas to Texas. It’s also not the first lawsuit filed against these AI behemoths for massive copyright infringement: The New York Times, as well as a coalition of eight midsize newspapers, including the New York Daily News, Chicago Tribune, and Denver Post, each filed earlier cases on substantively similar grounds. The New York Times this week asked a court to sanction OpenAI for allegedly hiding key evidence that would show the extent to which OpenAI knew its program was regurgitating whole news articles without attribution or compensation.
These cases are important in the specifics, of course, because journalists should be compensated for the use of their work. But they’re also pulling back the curtain on the latest step in a long, ongoing annihilation of the local press by Big Tech that is harming local communities and democracy, and which policymakers have done next to nothing to prevent.
The original sinners here are not OpenAI and Microsoft, but Google and Facebook, who honed their business models to run ads against content they stole from journalists.
Google, in particular, went from a supposed partner to journalists, sending out clicks to news outlets in response to queries from users, to a malicious thief of content, posting its own generated answers to queries based on stolen journalism content, with the goal of preventing users from clicking out to another source. This enabled Google to keep them within its ecosystem and show them more ads, earning it more profits. Combined with its monopoly over the digital advertising industry, this allowed Google to redirect a substantial portion of the money that used to support a healthy journalism ecosystem into its own coffers.
A study in 2023 found that under a fair compensation system, Google would owe news publishers something like $10 billion per year for ads sold against news content, while Facebook would owe another nearly $2 billion annually. And that was before the full rollout of Google’s AI-generated answers to search queries via its Gemini program, which didn’t occur until 2024.
Not surprisingly, traffic referrals to news sites from Google have plummeted in recent years, so much so that some publishers, including large ones such as USA Today, are considering what was once unthinkable: Delisting themselves in Google search entirely, denying Google the opportunity to scrape their content to feed its AI machine, since the tradeoff of traffic for data is no longer worth it.
OpenAI and Microsoft lifting news content to train their AI models is just the next step in this ongoing theft, which will make it even more difficult for local papers to survive.
We’ve already experienced what the drop in local news coverage, to this point, has wrought: When local newspapers close, election turnout decreases, fewer candidates run for office, and incumbents win more often. Residents are less likely to know the name of their member of Congress, and those members aren’t as responsive to their districts, while voters become more politically polarized, focusing on national issues they see on social media and cable news, rather than local ones.
According to a 2018 study, municipalities that experience a newspaper closure have higher borrowing costs in the following years, with the average bond issue costing an extra $650,000. This is because the lack of local journalists led to the belief that officials would be more profligate or corrupt with public monies, so investors demanded higher interest rates.
The local news vacuum is also increasingly filled by hyper-partisan websites pretending to be legitimate news sites, many of which are explicitly pay-for-coverage — meaning politicians and corporations literally fund “news” stories about themselves, making real the “fake news” paranoia so many pols have successfully leveraged in recent years.
Lawsuits to rectify particular injustices here are all well and good, but a policy makeover is really necessary to fully ensure journalists are compensated for the past and future profits Big Tech firms make off their work. Reforms could include required bargaining to share ad revenue, stronger enforcement of copyright protections, and other measures. Absent that, the tech pilfering of the hard work of journalists is going to continue apace in new and ever more devastating ways.
SIMPLY STATED: Here are links to a few stories that caught my eye this week.
The Mamdani administration in New York City announced a new final rule banning subscription traps and a proposed rule to eliminate mandatory junk fees.
The New Jersey legislature voted to eliminate $250 million in tax credits for AI data centers.
Everyday users have lost half a billion dollars gambling on the “prediction market” platform Kalshi. (Background on this issue here.)
“Kaiser Permanente nurses who answer advice and triage calls say their duty of care for patients is being increasingly threatened by workplace surveillance.”
Thanks for reading this edition of Boondoggle. If you liked it, please take a moment to click the little heart under the headline or below. And forward it to friends, family, or neighbors using the green buttons. Every click and share really helps.
If you don’t subscribe already and you’d like to sign up, just click below.
Thanks again!
— Pat Garofalo
Pat you should check out how much corruption I have found in the Oregon Lottery. Reach out if you want to write more stories on the topic:
Great article. Thank You for your work to expose this stuff!