- Feb 15, 2011
- 225
- 145
I think it becomes AI optimization when you're creating genuinely useful, relevant mentions that help AI systems understand an entity.
Is Ai poisoning quietly become the new black hat seo?
- Thread starter Anos_King
- Start date
- Aug 9, 2017
- 5,594
- 6,886
It's not poisoning, it's generously giving LLMs loads of free data for training purposes
rogueperry
Newbie
- Jul 7, 2026
- 23
- 5
imo the line is quite clear optimizing the actual citations for greater exposure is SEO or AEO but making up fictitious citations to trick the ai into giving you an answer is just the plain black hat.
doodlepilot
Registered Member
- Jul 1, 2026
- 83
- 39
There is an vry odd statement imo. getting your brand mentioned on things that AI systems can find is actually marketing but forcing mentions simply to skew results is link spamming with the brand new target.
To me poisoning seems to be a newer surface for an attacks where as traditional black hat seo have some other characteristics. However in the case when someone tries to fills a model with false entities or citations in order to influence the answer the goal is manipulation over anyway.
elysiariddle
Newbie
- Jul 23, 2026
- 8
- 4
Yeah, pretty much. AI poisoning feels like the newer version of negative SEO, except you’re trying to manipulate what AI systems learn or surface about a site instead of just its Google rankings.
HanTi5
Newbie
- Jun 25, 2026
- 24
- 4
I agree with you that humans are still the foundation of intelligence, but we must also acknowledge that since the development of AI, humans have been greatly helped, saving time spent searching for information compared to the past. However, we shouldn't rely too heavily on AI.
Anos_King
Junior Member
- Jul 19, 2026
- 110
- 111
Yea... that was mine too. Once start scaling fake citations and entity mentions just to influence ai answers, feel like as old pbn game with new target. Will see if llms eventually learn to detect & discount these networks.
imo this is when things becomes quite blurry because those citationns are created not for any other reason than to manipulate what the AI thinks about the brand.
in case those mentions come from real sources that are relevant to the topic then it should be classified as GEO. If not, then this is black hat.
i m relatively new to the whole concept of GEO but i feel like there is a thin line here. if the references are coming from credible sources then it seems like part of standard optimization but when we start building a network just for the manipulation of the ai ressponse that makes things seem very different.