If you were wondering if Google would be able to outsmart artificial intelligence, the early answers are certainly not looking good. Anecdotally, bot sites that have existed long before generative AI are likely going to embrace the new wave of technology to speed up their plagiarism, thus more robustly fooling Google into thinking that they are actual news sites, leading to Google promoting these stolen articles through the wildly popular and widely used Google News.
Anyone who has had to deal with Google Publisher services knows that the company has long ago abandoned its Publisher “Partners” when it comes to reporting false news sites, plagiarism, or low-value content. Google has been promoting sites “too small to have their own advertising department” artificially for years.
We’ve long reported when we’ve seen news sites copy-paste our work. Unfortunately, those sites often stay live for weeks, if not in perpetuity. Not only do we use the almost totally useless tools Google provides, but we also ping executives.
With AI and large language models (LLMs), it has gotten a lot harder, and Google, after years of letting its News product degrade for easy AdSense money, is now ripe for exploitation by even easily identified scrapers. As LLMs improve, they will become harder to identify. But what chance does Google have when it can’t even identify the low-value content it has been promoting for years?
Today’s example:
Rivian, the EV truck maker, released a software update last night that bricked many of its user’s infotainment systems. I reported on the outage early this morning for 9to5Google sister site Electrek. Within minutes, two separate AI sites had run my post through an AI generator and spit out similar articles.
Google News promotes those articles, sometimes ahead of the original:
The sites above use AI to rip off legitimate news articles, and not just from Electrek. The sites above also use Google’s advertising engine to monetize clicks over (my) scraped work.
The AI “author” on the OPP site has produced many paginated pages of scraped and AI-ed stories in the hour since my scraped article appeared on Google News. Each has a high-quality but easily identified AI-produced image attached to the news item. That would be an incredibly low-hanging signal for Google to follow if they indeed wanted to stem the tide of AI-produced content.
If I hadn’t already spent many hours uselessly reporting these sites, I would do that again.
But in reality, Google is just broken, and I’m not sure it is up to fixing itself at the cost of losing ad revenue.
FTC: We use income earning auto affiliate links. More.
Comments