I have spent some time denouncing Google’s excesses in Google Search, specifically those created with AI Overviews, which are already causing so many layoffs and media outlet closures all over the world. It is obvious to everyone that Google is a de facto monopoly that finally has a decent competitor (OpenAI and Anthropic afterwards) and perhaps that is why it is acting in a rather reprehensible way. In this context, I sometimes come across some AI Overviews that are quite outrageous.
Before continuing, let’s make clear what AI Overviews are. They are summaries generated by artificial intelligence (AI) that Google displays directly on the results page, usually in the best position, to directly answer a query. This reduces the possibility that the user will visit the source of the information, or in other words, the website that created the content.
Well, this week I updated my analysis of the Opennemas CMS and was looking at how it had ranked. When searching for “opiniones opennemas” I came across the following aberration.
I’ll summarize the situation for you: I work half a week on the article, publish it and Google likes it so much that it summarizes it and offers it directly to the reader. The opinions I collect in my article are real, newspapers that have contacted me and it is information that practically only I have. The entire AI Overview offers indirect links to my website, obviously, because I am the only one who has that information. So, you work hard, create quality content just as Google itself asks, only to end up being plundered by its AI. It is absolutely outrageous.
It is not the first time this has happened to me, nor will it be the last. On another occasion, I was the first to upload the top 10 GfK DAM data for media outlets and I found that Google was offering the article’s key data as a featured snippet directly, thereby clearly denying me the click:
The situation is a true aberration. And this comes on top of previous problems. Google is also testing replacing the headlines of media outlet news stories without consent, it flouts the AI anti scraping guidelines in robots.txt and a long etcetera.
AI Overviews, moreover, do not comply with Google’s own spam policies. Over time, the situation has improved (it no longer recommends, for example, eating pizza with glue), but, according to one study, 10% of AI Overviews still show errors, which means 1.369 billion errors every day and 57 million every hour.
But on top of that, taking my article and building a long AI Overview out of it is illegal. At no point have I given my consent for anyone to train on my content, nor for my content to be summarized, much less for it to be used against my interests, which is already the last straw. This far exceeds the right to quote.
But wait, the grotesque spectacle continues: another study recently noted that Google mostly cites itself in AI Overviews, mainly YouTube. And the same thing is starting to happen in Google Discover. It favors itself!
Fortunately, justice is moving forward, although slowly. Recently, the spanish CNMC has taken action on the matter (ENP Program) and Google has agreed to pay for snippets that go beyond the right to quote. Well, as for that “has agreed”, not literally like that: in practice, they are telling all publishers by default that they have no right to receive anything until they prove they are a media outlet under the law.
Because of this issue, I am having all kinds of disputes with them by email. I will publish something about it soon. Right now, I will only say that I am glad they respect the law to the letter when it comes to ENP, but I advise them not to do so only when it benefits them, but ALWAYS.
* Original article written in Spanish, translated with AI and reviewed in English by Jorge Mediavilla.
