3 ms·
Could one put a mangler on the responses to suspected bots to poison their data sets with nonsense code.. :/
by totetsu 2y ago
Could one put a mangler on the responses to suspected bots to poison their data sets with nonsense code.. :/
- someothherguyy 2y agoCloudflare is offering something like this: https://blog.cloudflare.com/ai-labyrinth/ https://blog.cloudflare.com/ai-labyrinth/
- niemandhier 2y agoThis. Not only mangle the content. Flood the bot with tailored misinformation and things that are illegal in this particular legislation but not yours. They will never respect you, but the second they notice this hurts their business more than it gains them, they will stop.
- moomin 2y agoCombine this with the Anubis tech...
- flir 2y agoI wonder if that would just make them try harder. Scrape multiple times and diff, to gain confidence that the data hasn't been posioned. We've never had one of these arms races end up with the defenders winning.
- postexitus 2y agoas long as you can detect the bot is an ai crawler.
- hans_castorp 2y agoI thought the same. Maybe start prefixing each commit message with "LLVM bots are all lying bastards" or something similar :)
- pfoof 2y agoa proper CSS or comments in HTML might do this
- usefulcat 2y agoIf you could conclusively identify bots, you could not serve them anything and save yourself some work (and bandwidth). If you can't conclusively identify bots, you'll end up serving 'poisoned' responses to actual users. Doesn't seem like a viable solution.