The AI Scam in Our Search Engines and Browsers: It’s Time to Put a Stop to This!
Today I’m angry! Because enough is enough! I’ve got a clear message for everybody in tech: when exactly did you decide we needed AI in our browsers and our search engines? When did somebody decide people can’t read anymore? That three lines cranked out by some obscure algorithm could replace the work of someone who spent hours building quality content?
Nobody asked for this bullshit! And you did it anyway! AI everywhere in our engines and our browsers. Usually at the expense of our privacy. And usually for features that are pure gimmick. You did it without asking and without warning, like it went without saying. Well, no! It doesn’t go without saying. It’s flat-out unacceptable! And that’s what I’m calling out in this article.
I’m also bringing radical solutions so we can defend ourselves against this assault. And I mean that word! Yes, it’s an assault, on web users and content creators alike. So as far as I’m concerned, you attack me, don’t count on me turning the other cheek. So listen up, tech people: without users and without content creators, you’re nothing! Absolutely nothing! You owe us everything and we owe you not one damn thing! Now that the message has landed, let’s get into it.
Google, Bing and now GPT: how American companies rig search results in their own favor
From Google and Bing, honestly, nobody’s surprised. Those two have always played the same game: pushing the sites that pay to show up and burying independent sites with no ad budget. That’s their business model, censorship that won’t say its name, and we’ve known it for years.
Their algorithms aren’t neutral, they serve a racket. And let’s not even start on the sites that bother libertarian ideology. Sites that go after Big Tech, anticapitalist sites, alternative media asking the real questions… Those get shoved so far down the results they barely exist anymore. In the end, why risk outraging the public with official censorship when you can just cap the audience? The question stands.
As if that weren’t enough, GPT Search and generative AI plugged straight into the engines take us one step further, and a much scarier one. They don’t just slant the order of the results anymore, they filter, summarize and reinterpret the content itself before it ever reaches you. A machine trained by American companies, with their biases and their interests front and center. From now on, if you don’t put in the effort, you’re in a total fog. The result: tech multinationals hand you on a platter whatever they’ve decided you should or shouldn’t know. That crude manipulation leads us straight to the death of the independent web.
Web search is a commons, not a market!
Web search has to be a public service! Yeah, somebody’s going to tell me that’s a communist idea. Well, I own it and we don’t give a damn! Because the real question is this: who should own the front door to the web?
Richard Stallman, Aaron Swartz, Tim Berners-Lee… None of those network pioneers built the foundations of the free web so three American corporations could grab them and turn them into a storefront. Berners-Lee invented the web and gave it to humanity. Swartz fought his whole life for free access to information, and it killed him. Stallman has been screaming for forty years that free software isn’t an option, it’s a political necessity.
And we let Google become the only front door to the web. We let a private, ad-funded company decide what exists and what doesn’t exist on the internet. That’s the real scandal! So this isn’t a completely idiotic political question, it’s a question of democratic survival.
Ecosia, Firefox, Brave, DuckDuckGo: we really did like you, but…
The part that really hurts in all this is how the so-called alternative engines and browsers behaved. Google and Bing, we crossed those off the list a long time ago. But you others, you were our alternatives. You carried the values of the free web, of privacy, of respect for the user. We could recommend you to our readers, our friends and our families, and be happy about it.
But you did the exact thing we were running away from with Big Tech. Exactly the same thing! AI summaries here, a built-in assistant there, an AI panel popping open that nobody asked for… Firefox shoving Microsoft’s Copilot into its sidebar. Ecosia summarizing your articles for you while it plants trees. Brave sticking Leo AI in the browser. DuckDuckGo with its AI Answers on by default… That list makes me want to throw up!
We’re not going to war with you over it. Credit where it’s due, you help a lot of people find NovaFuture. Honestly, thank you for that. But between friends, here’s the truth: you blew a historic opportunity! You had one card nobody else could play, being flat-out different, flat-out human, flat-out on the side of users and flat-out on the side of content creators. Instead you chose to follow the herd like sheep. So we’re just hoping this was a bump in the road and you’re all going to get a grip. Right?
AI summaries in search engines are a declaration of war on content creators
So what is an AI summary in a search engine, exactly? Here goes: a machine takes somebody’s work, digests it, spits out three lines and puts them in front of you so you never need to visit the original site. The content creator never sees your visit. Never sees anything come back from the work. Simply because he doesn’t exist anymore. So his ranking collapses, his visibility evaporates and his hours of work go in the trash. As for the AI, it helps itself without asking permission and without giving anything back.
A NovaFuture article isn’t some throwaway thing. It’s hours of research, writing, fact-checking and formatting. It’s written with two kinds of love: the love of writing and the love of sharing. Every long article is long for three reasons. Because the subject deserves it! Because the ins and outs don’t fit in three lines. Because cutting it short means betraying it. And when a machine decides for you that you don’t need to read the original, it isn’t doing you a favor. It’s robbing you of an experience, a shade of meaning, a shared feeling and a complete thought.
And on opinion pieces it’s even worse! An AI summary on a political, social or philosophical subject will always drift toward whatever reassures you, whatever is consensual, whatever bothers nobody… It’s just an automated echo chamber. Never mind that a summary is stupidly easy to manipulate. You’ll never know, because you didn’t read the original and you trusted three lines from a machine whose biases and real intentions you know nothing about. So here’s the summary: everybody needs to learn how to read again, goddammit!
How to block AI scrapers to protect your site and take back control as a user
Let’s start with the technical reality. The big AI companies don’t give one single damn about your robots.txt file. It’s a convention, not a law. Same as taping a “keep out” sign on an open door. Blocks in your .htaccess are more serious, but they get worked around more and more easily. The bots rotate user agents and IPs to get past them. Some go as far as passing themselves off as web browsers. It’s an endless cat and mouse game.
The real fix is dynamic blocking at the server level. A smart system that watches request behavior in real time, spots the patterns that give scrapers away and blocks them automatically before they’ve swallowed anything. It works. Sadly it’s only for people with a VPS or a dedicated server who really know what they’re doing. On shared hosting you’re screwed, you just don’t get access to that layer of configuration.
So here’s a call to developers, because the web badly needs simple, accessible tools that work. CMS plugins, for instance, that do dynamic blocking without the webmaster needing a PhD in sysadmin. It’s doable and not necessarily complicated. So, developers who honor us with your presence here, take note. We don’t have the time to throw ourselves 100% into a project like that. But we’re happy to help where we can if you need it. So far, everything we’ve seen says only algorithms can fight harmful algorithms effectively. Anything static doesn’t work, or at least doesn’t really work for long.
Now for users. If you haven’t done it yet, turn off AI summaries in the settings of every engine and browser where that’s possible. Switch to clean alternatives like Searx, Marginalia, Mojeek or Librewolf. And above all, don’t stay quiet… Make noise! On social media, on forums and by email straight to the people who ship this stuff. Say it plainly to Firefox, Ecosia, Brave and DuckDuckGo: you don’t want their AI. That you picked their tools precisely to escape this and you feel betrayed… Put on as much pressure as you can, because pressure works. Your silence only makes the problem bigger.
AI is just a tool. We’re not here to put software on trial.
The world isn’t black and white. We’d be dishonest claiming AI is pure evil. A tool is a tool. Everything depends on who’s holding the handle and why.
AI to speed up scientific research, to help doctors make sharper diagnoses, to improve weather forecasts and see climate disasters coming… We’re 100% for it. No hesitation. That’s exactly the kind of use that justifies the years of research and the billions poured into this technology. It’s useful, it’s concrete and it can make life better.
But AI to spy on people, to vacuum up creators’ content without asking them or paying them, to steal artists’ work and regurgitate it in their place, to summarize other people’s thinking and make everybody more passive and more dependent… That’s a hard no! And this harmful AI has to be fought right now, not in ten years when it’s buried even deeper in our habits and our tools.
The difference between the two? Simple. In the first case AI serves people. In the second, people serve the AI and the commercial interests of whoever controls it. This is a massive choice about what kind of society we want.
Thinking AI saves you time is a big illusion
Let’s be honest, because this matters. On paper AI should save time. In practice it’s a lot messier than that. If you use AI seriously, you can’t afford to trust it blindly. No exceptions: you check every claim. You cross-check every number. You verify every source. Because AI hallucinates, invents, mixes things up, oversimplifies… and does all of it with a confidence that makes any professional liar look like an amateur. And if you don’t check everything hard, your credibility goes to zero. So in the end, the time savings aren’t there at all.
And a new problem came with it. The web is filling up with AI-generated content at a terrifying speed. Articles, analyses, complete guides, in-depth studies… all of it churned out at industrial scale, with no thought, no expertise and no soul. So when you research something for work, now you have to watch out for that too. Sorting the real from the synthetic is an extra layer of work we didn’t have before. So the time you save on one side you lose on the other. And not many people are saying it.
What if we built our own search engine?
Yeah, we know what you’re thinking. Same diagnosis as always, but what do we actually do about it? Nothing, like usual? Well, funny you should ask. Here’s our answer:
What if we built our own independent search engine? A real one, 100% free and 100% open source. We’re not kidding! We’re talking about a real project, doable and genuinely useful. Not a utopia dreamed up by geeks with delusions of grandeur. A tool that serves users and content creators.
The concept I have in mind is a hybrid between a classic engine and a collaborative directory. For the interface, a totally stripped-down look. Just a search box and a submit button. Like Google in 1998, before it turned into the devil despite all its pretty promises. Remember? “Don’t be evil.” We all had a good laugh back then, because we got right away that it would go bad.
Let’s get serious about the technical side. The crawler is Scrapy in Python, solid, scalable and battle-tested. It crawls the web and rates every site as it goes, against clear and transparent criteria. The index lives in PostgreSQL. Simple, reliable and easy to audit.
Scoring is the heart of the project. Every site gets a grade based on weighted criteria. AdWords or Bing or any other ad network detected through DOM analysis? Not indexed. Trackers identified against the EasyList and EasyPrivacy lists? Not indexed. Mass AI-generated content flagged by a home-trained classifier or an API like GPTZero? Not indexed. Garbage already has a place to go and it won’t be this engine.
For the cases automation struggles with, scam sites, borderline content… we add a layer of collaborative human curation. A community that flags, approves, rejects… Not an opaque algorithm laying down the law in its corner. Humans doing the checking, out in the open. That gives you a hybrid between machine power and collective intelligence.
The backend could run on Flask or FastAPI, light, light and fast. With an open API so anybody can contribute, fork or improve it. One VPS is enough to start. That way the architecture stays modular so it can scale if the community grows. All of it self-hostable, decentralizable and free.
Is it easy? No! Is it impossible? Absolutely not! It just takes motivation, skills and infrastructure. Above all it takes people who are sick of taking it and would rather build. I’m sure that speaks to you 😉
Conclusion: speaking honestly to move forward
From the bottom of my heart, I spent several hours putting this article together. The only thing driving me was the thought that if it helps even one person feel less alone in thinking what they think, that would be a real satisfaction for me. This is NovaFuture, and sometimes we probably look like a bunch of Don Quixotes charging at windmills. I think that myself now and then. It passes fast, because at least we have the merit of not giving up, and the monsters we’re fighting are very real. And we’ll win in the end, because we have nothing to lose and nothing to win either, except our stolen freedom.
No AI will ever summarize that feeling for you. No AI will ever care about who you really are. No AI will ever tell you how to change the system. Only real human exchanges can move us toward something good. Basically I’m just a geek you could fairly call antisocial. Maybe that’s true, but this one time I really wanted to share what I actually think, with the grim mask of the guy typing code into his Linux console taken off. So thanks for reading all the way to an ending that’s only the start of a story still being written in the back alleys of the underground web. However many of us there are, all that counts is standing together on the idea that another world is possible. So, see you very soon for concrete action.
If you enjoyed this article,
support copyleft and open source
and help us bring our projects to life.
