##
Quoted in Ars Technica’s article on tarpits for AI crawlersI am quoted at length in a new article on Ars Technica (a website I have been reading for years and years) on an in-depth [article on tarpits for AI crawlers](
). But people who know my MO will already guess that even being quoted at length is only the tip of the iceberg and [Ashley Belanger](
) generously agreed for my to publish my full answers and her questions here.
So read [the article]( ) on AI tarpits and here’s the “interview” we did via email (questions in italic are Ashley’s “prompts”):
*1. What do you think is cool about Nepenthes? Any weaknesses?*
I like how Nepenthes changes a dynamic (whether it’s fully functional is another matter I’m gonna get to) that has been going on for months if not years: People feel helpless. AI companies scrape all there is without regards to technical means (robots.txt) or established law (copyright). Because the AI people (usually men) are our special boys, they get to do whatever they want and we can only accept it. Nepenthes says: NO.
It’s an expression of a demand for agency by the inhabitants of the internet and that’s what I think is cool. It’s a symbol that the story of there being no alternative or choice is false, regardless of how often AI companies and AI boosters tell it.
Nepenthes is probably not gonna be useful in the long run. First I don’t think that enough people would run it to make a meaningful difference in the huge datasets that AI is being trained on. Yes, it’s drivel, but there’s already enough in the data. Second I think that if it *did* become a problem, companies would find heuristics and detection algorithms to find Nepenthes instances and just drop the garbage text.
Nepenthes is (in my reading) more of a sociopolitical statement than really a technological solution (because the problem it’s trying to address isn’t purely technical, it’s social, political, legal, and needs way bigger levers). It’s a powerful symbol.
*2. What other defenses against AI training have interested you?*
Of course there’s Nightshade https://nightshade.cs.uchicago.edu/whatis.html and a few similar projects. All are interesting and especially if one’s business depends on one’s very specific and distinct art style those tools might make sense (if they could be applied retroactively). But all that has been taken isn’t protected by it. And while those systems are technically interesting I don’t think that a cat and mouse game of obfuscator vs. crawler is really how we as societies should spend our time in the long run – especially given that this does not at all address all the other harmful aspects of AI systems (environmental impact for example).
3. Nepenthes creator has told me he is firmly in the camp of wanting to watch AI burn and that he doesn’t care if Nepenthes is successful, he just wants to do anything to fight enshittification of the Internet. Any thoughts on why AI in particular is fueling this “burn-it-to-the-ground” energy?
I think AI is just the latest in a long row of technologies, it’s just the most aggressive one: Technologies that are not done *for us* but *to us*. The tech sector has been struggling to produce meaningful innovation in a while (I talked about that at last year’s Re:Publica conference in Berlin 
) so it keeps spitting out more and more things that people feel nobody asked for. Nobody asked for NFTs or the fucking Metaverse. Nobody asked for lying chatbots instead of getting to reach an actual support person that could solve your problems.
I think this feeling fuels the dynamic: A bit of disappointment in tech and the way it can’t seem to do interesting things anymore and a feeling of tech being built not for the good of anyone but VC investors and a few tech dudes. It feels a bit like the social contract that society and the tech sector/engineering have had (you build useful things and we’re okay with you being well off) has been cancled from one side. And that side now wants to have its toy eat the world. People feel threatened and want the threads to stop. And who can blame them given *points everywhere*.
*4. Nepenthes creator told me that Google’s and Amazon’s Claude crawlers have been trapped by the tool but OpenAI’s crawler has managed to escape. Critics have suggested big tech companies will defeat any tarpits like Nepenthes as they crop up. Who do you expect would be most affected by projects like Nepenthes?*
I think a few smaller crawlers might not have defenses up, but AI is just a numbers game that naturally centralizes. So I don’t see any relevant player being affected in the future.
*5. Any other comments on tarpits or similar efforts to trap web crawlers and poison AI models?*
Nepenthes isn’t alone, it has sparked followers that (like ) are sometimes even more aggressive, that try more aggressively to poison the data well for AI companies. And I think that while those systems also are basically just symbols it’s a great sign to see that people are challenging the notion that we all have to do AI now. Because we don’t. It’s a choice. A choice that mostly benefits monopolists.
We as people critical of AI shouldn’t believe those systems will solve the issue, but we should embrace the feeling of agency. Of getting back to a mode of action. And then we need to transfer that to political and social action.
Liked it? Take a second to support tante on Patreon!
[](https://www.patreon.com/tante?utm_content=post_button&utm_medium=patron_button_and_widgets_plugin&utm_campaign=78367&utm_term=&utm_source=https://tante.cc/2025/01/28/quoted-in-ars-technicas-article-on-tarpits-for-ai-crawlers/ )
[]( )This work is licensed under a [Creative Commons Attribution-ShareAlike 4.0 International License]( ).
Ars Technica
AI haters build tarpits to trap and trick AI scrapers that ignore robots.txt
Attackers explain how an anti-spam defense became an AI weapon.
Ars Technica
Author: Ashley Belanger
Ashley is a senior policy reporter for Ars Technica, dedicated to tracking social impacts of emerging policies and new technologies. She is a Chica...
Ars Technica
AI haters build tarpits to trap and trick AI scrapers that ignore robots.txt
Attackers explain how an anti-spam defense became an AI weapon.
Smashing Frames
"Empty Innovation" at Re:Publica 2024 #rp24
Under the motto “Who Cares” Re:Publica gathered a few thousand people in Berlin and I got to give a talk continuing the path I started ...
Cookie monster!
Deed - Attribution-ShareAlike 4.0 International - Creative Commons
Deed - Attribution-ShareAlike 4.0 International - Creative Commons
