I think I found a simple hacky way to mess with ai spambots

[Rubbereon]

And the funniest part? It takes almost no effort and I assume it does a decent job against ai bots with no drawbacks. Just do this:

Code

        <script>
        var link = document.createElement('link');
        link.rel = 'canonical';
        link.href = 'https://yourwebsite.org/index.html';
        document.head.appendChild(link);
        </script><!--The canonical link may need to be different than the webpage url, but the crawler should see this while the ai can't.-->
        <noscript>
        	<link rel="canonical" href="https://anotherwebsite.com/">
        </noscript><!--Where "anotherwebsite.com" is either a dead end (like google.com), a tarpit, a fountain that poisons the ai or absolute nonsense gibberish e.g. https://divinicide.net/blackpages/ that would basically do the same thing-->

I know many folks had problems with ai, so I am sharing this in the hope that it is helpful to anybody who runs a website by themselves.  

[Seechain]

Code

        <noscript>
        	<link rel="canonical" href="https://anotherwebsite.com/">
        </noscript><!--Where "anotherwebsite.com" is either a dead end (like google.com), a tarpit, a fountain that poisons the ai or absolute nonsense gibberish e.g. https://divinicide.net/blackpages/ that would basically do the same thing-->

Code

<noscript>
 <link rel="canonical" href="http://127.0.0.1:22">
</noscript>

This topic seems very interesting to me. Basically, Black SEO techniques are used for protection.

[Noah_S]

http://127.0.0.1:22

This address is restricted

This address uses a network port which is normally used for purposes other than Web browsing. Firefox has canceled the request for your protection.

My firefox looks like it wonders what I was even doing

[MediocreMetastasis]

You can punish bots that don't read your robots.txt file by placing landmines https://news.ycombinator.com/item?id=44670319'
It aint much but it does affect the web scraper a tiny bit.

[Seechain]

This address is restricted

This address uses a network port which is normally used for purposes other than Web browsing. Firefox has canceled the request for your protection.

My firefox looks like it wonders what I was even doing

I hadn't seen your reply. In the script flow, that part is only visible to the bot. It's a joke, kind of like telling the bot to go sniff its own butt.

-------------------------------------------------------------------------------------------------------------------------

I found this and wanted to share it.





Apparently, there are webmasters who are taking this quite seriously.

The website is this: https://tcrf.net

⚑ Moderators Note ⚑

Merged in some extra context from another thread the OP made duplicating this post. @Rubbereon I've directed discussion of this here since its your thread and you wrote a good reply below; however if you don't want this in your script thread let me know!

[Rubbereon]

I found this and wanted to share it.




Apparently, there are webmasters who are taking this quite seriously.

The website is this: https://tcrf.net

I was excited to learn about the anti-bot thingy but then I realized as I finished reading this what site this is on...

First off, the anti-claude protection is essentially closed-source, so no one else is using it and secondly, I generally don't regard tcrf favorably because xkeeper (the owner) is problematic, treats her own site like her own little kingdom that she fully owns and no one else can or will own ever (quadruple emphasis on "own") rather than a democratic archival site like say archive.org where they even offer torrents for anything, and she hates altruistic archival (or in other words, archiving for the sake of allowing others to access the content later) because she blocks ips nilly-willy. For example, she blocked archive.org, she blocked and ip banned anybody who tried to clone the wiki (which she blames it on the green bird site that shan't be named) and she has blocked possibly untold amount of other people for simply discovering her site becourse on top of everything else she has set up a script to ip ban indefinitely anyone with a specific url referrer and regardless of whether that is justified this is both irresponsible, ineffective (vpns and tor exist duh, she is possibly blocking innocents on dynamic ips while the people she wants blocked don't care and will bypass her script) and nuclear withoput any real benefits, and just overall she has rancid vibes, she feels exactly like Keffals to me. I actually blocked her via a moderation list on bluesky, no thanks.

This is the short version though, I could argue there is an entire iceberg on her because she gets into internet slapfights regularly and to my knowledge has tried to start drama with romhacking sites and forums for one reason or another. I hope you get the gist of it because I don't think Meloon wants me to write a 10 pages essay on her, that's not the point of Melonland and neither it is of this thread.

I don't know any sites that explicitly block ai llm bot requests with a funny splash screen, so sadly I can't contribute to this particular discussion, all I know is that most sites use cloudflare to block bot requests and there is again the fact you can use a tarpit, poisoned data, a dead end, pure javascript or css (which the ai bot can't parse) or a general purpose anti bot tool like anubis to mess with ai bots. Maybe I might do that once I open a serious interactive website like a forum, who knows.

Thing I can add though is I've learned about shieldfont recently which is a special font that shows readers real text, but to ai bots it is pure gibberish. I haven't used it because I want people to be able to copy all the text on my website and I'm worried it makes that impossible, but if you're a writer you may not want anybody to copy what you've written anyway. In addition to that, there is the moral implication that screen readers may or may not be able to read the text either so...


ETA before posting: Turns out that xkeeper personally blocked him by hand and fully acknowledges him on her bluesky, she might be stalking him to see if he posts anything new about her website. What must've happened is that Linode saw a surge of requests from his use of claude and reported them to xkeeper as a ddos attack which she too can see because she has a panel open with every request passing through including user agents and http referrers (proof enough that there's likely nothing installed serverside, although most people would normally rely on cloudflare, anubis or modsecurity) Yeah so I know you didn't mean any harm, but this is 100% not a good example of an anti-ai stance because it's neither truly genuine (as mentioned earlier it is comorbid to wanting to ban any and all automated content retrieval, see again the archive.org ban) but also the system may not exist whatsoever and the bans may be handed out by hand.

At best it's misleading as you can technically use linode and they should report back to you any odd behaviors like llm bot spam or genuine ddos attacks and then act of them.