I’ve been thinking lately about AI poisoning. I want to make this post a thread for suggestions about which tools we can use to fight back. Bear in mind that I’m only talking about poisoning tools, not blocklists or bot-detection utilities. Here’s a short list of what I have so far:
Website protection tarpits for text poisoning (fighting web scrapers):
- Nepenthes was one of the first. In the words of its author, “I’m just fed up, and you know what? Let’s fight back, even if it’s not successful. Be indigestible. Grow spikes.” Nepenthes’s demo page is here.
- Iocaine is the best tarpit I know of right now. It’s very well-designed and performant. Version 3.0x is complete, and the fourth version is a work in progress. It’s made by Gergely Nagy, who goes by the pseudonym “algernon”, after the character by the same name from Flowers for Algernon, by Daniel Keyes. Iocaine’s demo page is here.
Image poisoning tools:
- Nightshade is an image poisoning tool developed by the University of Chicago. Nightshade is a part of the Glaze project, which also includes WebGlaze and Glaze itself.
I just read an article about someone analysing browser fingerprinting, and he found some very interesting things that e.g. unpacked into a list of keywords that made everything AI balk because of hardwired security rules. The list of keywords included traps for American AI like “Antrax”, “Bioweapons”, etc, and some for Chinese engines like “Uigur” and “1989 Tiananmen”.
It is not exactly poisoning, but a nice deterrent has it’s merits, too. And to have it encoded wastes a bunch of cycles and tokens extra on the other side, too.