--- title: "SEO Hacks and AI Poisoning: Can You Really Trick AI?" description: "Every SEO hack worked until Google said no. Now AI can be poisoned with a few planted words, Google calls it spam, and I think honesty finally wins." url: https://wonderworksdesign.net/authors/colin-rooney/predictions/you-cant-bullshit-ai language: en updated: 2026-10-16 publisher: Wonder Works (Wonder Works Design Ltd), Blagoevgrad, Bulgaria --- [Home](https://wonderworksdesign.net/)[Colin Rooney](https://wonderworksdesign.net/authors/colin-rooney)[Colin's Predictions](https://wonderworksdesign.net/authors/colin-rooney/predictions) 04Called on **16 Oct 2026**Still open # You can't bullshit AI? [Colin Rooney](https://wonderworksdesign.net/authors/colin-rooney)Co-founder, Wonder Works · my opinion, not Wonder Works advice ![Colin Rooney with his arms folded and a sceptical look](https://wonderworksdesign.net/media/colin-arms-folded-navy.webp) My social media was full of SEO "hacks" for years, and it still is, a bit less now, but there were a few in my feed today. Do I do it? Yes and no. We practise SEO, and there are tricks to it, but we don't lie, it's just not worth the risk. We don't keyword stuff, because we don't need to, and you shouldn't either. For me a keyword is just a way into the topic people are talking about, so if it's relevant, and only if it's relevant, we'll talk around that subject on our site and on social. Not for ranking, because honestly I don't see ranking as being as valuable as it was, and I'm sure millions will disagree, but for relevance. ## Right now you can bullshit AI a bit Every one of those hacks worked once. Keyword stuffing, crap backlinks and fake reviews all worked until Google said no. Now the new ones are about AI. [A study from Cornell Tech this year](https://www.truelogic.com.ph/?p=90927) found that thirteen words planted on a busy forum page could slip a brand into AI research reports in up to half the tests. So yes, right now you can bullshit AI a bit. But in June [Google started treating attempts to manipulate its AI answers as spam](https://www.truelogic.com.ph/?p=90927), the same way it went after keyword stuffing, and we've all seen how that film ends. It's just not worth the risk. ## Relevance and being honest and real is what works It used to be that the ones who showed up in search, and more importantly on social media, were the ones with the biggest budget and the bigger brand. I don't believe that anymore. Relevance and being honest and real is what works, and we've seen it. One of our clients had 406 followers when a post we did for them went past 160,000 views, with not a penny spent on ads, and three weeks later the following had nearly tripled. ## There is no Harry Potter moment And it takes time. Nothing is fast, and there is no Harry Potter moment that changes everything. Some of it is luck, but most of it is time and consistency. Telling the truth is, for me, the only strategy, but be prepared to wait. Sometimes you just hit the sweet spot, and again some luck is involved, because I don't care what anyone tells you, there is no playbook for a viral post or an AI citation, definitely not one thing anyway. It's consistency and time, and again, for me, honesty. > Telling the truth is, for me, the only strategy, but be prepared to wait. ## Why not be the one that tells them? I have a client who is reluctant to share things like recipes, or the way they do things. It's their business, of course, and they should be able to say and not say whatever they want, but mate, information is sooooo free for everyone now. They will find what you don't want them to know, so why not be the one that tells them? If you don't say it, AI will fill the gap with whatever someone else said about you. But hey, that's just me! I grew up in a world where nice guys came last. Man, I hope this changes, and I genuinely believe this is the beginning of it. ## Updates, newest first 16 Oct 2026The original call This is the prediction as I made it that day. When something changes, I add a dated update here and leave the original as it is. ### In this piece 1. Right now you can bullshit AI a bit 2. Relevance and being honest and real is what works 3. There is no Harry Potter moment 4. Why not be the one that tells them? ### How I mark these - Happening - Still open - Proved wrong ### Linked predictions - [Google won't be Kodak](https://wonderworksdesign.net/authors/colin-rooney/predictions/google-wont-be-kodak) - [Long live the AI wars](https://wonderworksdesign.net/authors/colin-rooney/predictions/long-live-the-ai-wars) - [“I don't need a website, that's what ChatGPT is for”](https://wonderworksdesign.net/authors/colin-rooney/predictions/i-dont-need-a-website-thats-what-chatgpt-is-for) - [SEO is named after something nobody does anymore](https://wonderworksdesign.net/authors/colin-rooney/predictions/seo-named-after-something-nobody-does) ### All of them - [← Back to the logbook](https://wonderworksdesign.net/authors/colin-rooney/predictions)