In February 2009, I announced the launch of the “3 Count” column, a 4X per week column covering copyright and copyright-related news.
The goal of the column was two-fold. First, I wanted a place where I could briefly touch on and cover copyright news without the overhead of doing a full post. Every week there are many copyright news stories that, while important, I don’t have much to add to, so this is a great way to cover the news and move on.
Second, and perhaps most important, it was a meant to showcase ethical and legal aggregation. The format features clear links to the source material, rewritten copy and, in more recent versions, clear bylines to both the publication and the reporter(s) that wrote the article. The format was devised through consultations with reporters, SEO experts and others involved in the process.
The column has gone on to be a staple of this site, and, since that initial publication, I’ve put out over 2,600 3 Count columns, each highlighting three links that I think readers might be interested in. (Including one misguided April Fools’ joke edition)
However, this had an interesting impact on my daily work routine. Most mornings (or late nights), one of my tasks is to pour through Google News and find links for the 3 Count column. It’s a ritual that I’ve done well over 2,600 mornings over the past 14 years.
For most of that run, it’s actually been a relatively pleasant part of my mornings. Something to go with my first coffee as I finish waking up. But, in more recent years, it’s become increasingly challenging.
Where once I would routinely get 10 or more pages of results, most of which were actually useful, now I am lucky if I get 3 full pages of results each day. To make matters worse, once you filter out repeat stories, AI-generated garbage and opinion pieces/press releases presented as news, I’m lucky if there are three stories to be found.
Where once Google News was a haven for international and local news sources that I never would have found otherwise, now it’s overrun with AI knock offs, unethical human aggregation and stories that aren’t actually news. Virtually no human-written news is allowed.
I am not the only one who has noticed this. Yesterday, Seth Weintraub at 9 to 5 Google reported very much the same thing, saying that Google News has become overrun with clear AI garbage, though he acknowledges that bot sites and other low-quality content has been running roughshod over Google News for years.
In a word, Google News is dying. It’s being eaten from the inside out. While AI may be what drives the final nail in the coffin, the cancer has been growing long before the rise of LLMs. To make matters worse, Google likely could stop or at least slow this decay, but they have no real motivation to do so.
In fact, it’s likely in their best interest to just let it rot.
Google’s Copyright Ground Zero
From Google’s perspective, Google News has been a nearly two-decade long copyright headache that shows no sign of ending or slowing down.
All the way back in 2006, a group of Belgian newspapers filed a lawsuit against Google, alleging that the search engine had no right to use their links and snippets without payment or permission. They would eventually win that case, in 2011, kicking off a retaliatory block by Google of those papers.
However, it was just the first salvo in a battle that would be fought in both courts and legislatures all over the world. The result of that fight was that, in March 2019, the European Union passed a copyright reform bill that required search engines and aggregators to pay for headlines and snippets. Australia passed a similar act in 2021 and both the United States and Canada have similar bills in progress as of last month.
This has forced Google to the negotiating table with news publishers, with Google saying that they pay some 450 publications for access to their content. This has also prompted Google to launch Google News Showcase, which it claims is an effort to “help people get the quality news and information that matters to them” through a $1 billion investment in news organizations.
Though these issues could, in theory, impact all of Google’s search products, the focus has been heavily on Google News and its relationship with publishers. According to Google, they don’t make money from Google News and that news websites are a “small slice” of the internet and that news queries account for less than 2% of total Google search queries.
While those statements may well be true, news publishers constitute a disproportionate percentage of the quality and informative pages on the internet. Though their numbers may be few, especially today, they are a crucial part of the internet for anyone who wants accurate and well-crafted content, especially on complicated subjects.
But, as publishers have been fighting for payment for their content, other groups have been more than happy to let Google hoover up their sites for free: Creators of bot websites and websites pushing specific agendas.
To that end, Google has been more than happy to welcome them, especially if they support Google’s best interest in other ways.
The Rise of the (Non-AI) Bots
Back in April 2010, my friend Patrick O’Keefe and I looked at a site named Global Grind, which was scraping content from various blogs, including O’Keefe’s, and republishing them onto Google News.
Note: At the time, Global Grind said that the scraping was due to a “legacy” system that was still publishing content, and the scraping stopped shortly after publication of our articles. The site is active today, though it is unclear if it has any connection with the version from 2010.
At the time, this was unheard of, and it was widely assumed that the only reason Global Grind was accepted into Google News was its association with the celebrity businessman Russell Simmons.
That’s because, back then, admission to Google News was severely restricted. While scraping and republishing was an all-too-common tactic, Google News was, at least up to that point, a respite from that behavior.
However, since then, the standards for getting into Google News have been significantly lowered, with more and more sites being accepted. Though many of those sites did and still do high-quality work, a significant percentage were producing, by any reasonable metric, garbage.
This includes scrapers, but those numbers have been overshadowed by sites that rewrite news content with little to no attribution. Some of these sites would do it with human labor, having authors “write” a huge number of articles per day, and others would just use bots, such as article spinners, to “rewrite” the content and make it appear original.
Either way, legitimate sites could expect to see “rewritten” copies of their articles within hours or minutes of theirs posting. However, for a time, it at least seems as if Google was fighting these sites, often removing them within days or weeks of reporting, as they did with the Panda/Farmer updates of February 2012.
But then came the next big leap in “content generation” as AI took an existing issue and took things to a whole new level.
AI and Perverse Incentives
Though generative AIs have been around for many years, it was the public launch of ChatGPT in November 2022 that fired the starting pistol on public use of large language models and generative AI systems.
For spammers, AI was simply a new tool that they could use to more quickly and more effectively pump out low-quality content. The rise of generative AI spam was one of the most predictable consequences of this shift.
However, where Google put up a facade of fighting other kinds of spam, they’ve completely stopped pretending when it comes to AI spam, even changing their content guidelines to allow the use of AI.
The reason is fairly simple. Google, through Bard, is one of the largest investors in AI and one of the leaders in the space. It also doesn’t hurt that, where most large publishers use their own advertising system, smaller AI-powered sites often use Google Adsense.
In short, if Google were to take serious action against AI spammers, it would both harm their investments into AI and their ad revenue. From a purely practical standpoint, there’s no reason for Google to step in, at least not until it hurts the user experience so much that users begin to leave.
But where would those visitors go? Bing News has many of the same quality issues, and their owner, Microsoft, is also a heavy investor in AI. There’s no obvious competitor, at least not one with the scale and scope that can seriously challenge Google.
Sandwiched between publishers that they’ve only begun to begrudgingly pay and their own business and legal interests in AI, Google has every reason to not solve this problem.
In short, if Google News is truly rotting and dying, it’s because Google has no reason to stop it.
As I said in the beginning, Google News has been a staple of my life for over 14 years. Multiple times per week, I’ve been performing the same or similar queries, and I’ve had a unique vantage point on how the service has changed.
It seems that, as the battles with publishers heated up, that Google loosened its criteria for inclusion and welcomed sites that were lower quality, more biased and, in many cases, not even news sites.
Though this erosion has been taking place for some time, the rise of AI has greatly accelerated it. Now, not only do the spammers have a new tool that enables them to pump out more “content” than ever, it’s a tool that Google has every incentive to ignore.
In that regard, the death of Google News is a microcosm for what is going on with the rest of the internet. It’s a warning to what happens if AI goes unchecked.
No matter how you feel about AI broadly, Google News is an example of what happens when you let the worst practices of AI go unchecked. Nothing in these garbage sites point to the promise or potential of AI, only how it can be used to shove out human creators by spitting out large quantities of lower-quality work.
Whether you believe AI is the future or hate what AI represents, this should worry you.
However, it clearly doesn’t worry Google and that, in the end, is the biggest problem of all.