LLMs Are Malleable: What Black Hat Tactics from 2010 Mean for AI Search

LLMs are deeply malleable, and right now, they are completely susceptible to the spam tactics that kind of worked pretty well for Black Hat SEO ten years ago.

You can go on and get spammy listicles, you can manipulate what an LLM says by Reddit threads. I shared that one Reddit story about how this one person had tried to launch a complete defamation of their competitor through Reddit, and were deeply successful doing that. It really kind of ruined their reputation on Google, on Reddit, and in LLMs because of the importance of those forums.

The spammy tactics work right now for LLMs.

The Black Box Problem

We don't have any concrete ranking factors from LLMs. It's a little bit of a black box as to how they ingest the information, how they output their citations, how they decide to share different sources based on the personalization of my search versus someone else's search.

That lack of transparency creates vulnerability. When we don't know the rules, we can't protect against manipulation. And right now, bad actors are absolutely exploiting that vulnerability.

Why Forums Are the New Battleground

Forums are really gaining in popularity in traditional search results. Almost any Reddit thread that pops up in the top ten on traditional searches are two to three years old. They're not yesterday. They're not conversations that happened a month from now.

And LLMs are ingesting that forum content as training data, as source material, as supposedly authoritative voices of "what real people think."

Which means if you can manipulate the forums, you can manipulate the LLMs. You can plant negative information about competitors. You can astroturf positive sentiment about your own products. You can create entirely fictional narratives that AI will then repeat as fact.

This is actually happening. Not in some theoretical future—right now.

What Happened to Content Quality Signals?

In traditional SEO, Google developed increasingly sophisticated ways to detect spam. They looked at link patterns, content quality signals, user behavior, domain authority, all of these things that made it harder and harder to game the system with cheap tricks.

LLMs don't have that sophistication yet. Or if they do, they're not applying it consistently.

They'll cite a Reddit thread with three upvotes as readily as they'll cite a peer-reviewed journal. They'll pull from sources with no verification of accuracy. They'll repeat information that's been planted specifically to manipulate them.

The Reputation Attack Vector

Here's what keeps me up at night: it's not just about promoting yourself dishonestly. It's about attacking others.

That Reddit defamation example isn't an isolated incident. If someone wants to damage your brand's reputation in AI search, they can. They can create forum posts claiming your product doesn't work, your company has terrible customer service, your leadership is unethical—whatever narrative they want to plant.

And if they do it across enough forums, in enough places, with enough seeming authenticity, LLMs will start incorporating that into their responses when people ask about your brand.

You won't even know it's happening until someone mentions "I asked ChatGPT about your company and it said..."

What This Means for Brand Protection

Everything we say is going to matter, and it's going to matter even more just going forward. We have to be so intentional about what we put out there for ourselves, for our clients.

Not that this should put us into paralysis of "now we can't say anything because everything matters so much," but being really thoughtful about the narrative we're crafting and everything fitting into what that narrative is.

We also need to be monitoring. Not in the traditional SEO monitoring way, but actively checking what LLMs are saying about our brands and our clients' brands. Testing different queries. Seeing what sources they're citing. Looking for patterns of misinformation.

The Evolution We're Waiting For

Search engines took years to develop good spam detection. LLMs are going to go through that same evolution, but we're in the early messy phase right now.

Eventually, there will be better verification systems. Better source quality signals. Better detection of coordinated manipulation campaigns. But we're not there yet.

And in the meantime, the Wild West is open for business.

Where Quality Still Wins

Here's the thing, though: quality is still going to win the war. It doesn't necessarily matter how we get to that output, but if the output is quality—in terms of the information architecture, in terms of the technical soundness, in terms of what we say, how we show up, where we show up—if that is high quality, we are going to win.

We've seen that, and we can prove that we can still do that at scale.

But we can't be naive about the fact that others are playing dirty. We can't assume that just doing good work is enough protection against manipulation.

The Strategic Response

We need to be present in forums and communities where conversations about our brands and industries are happening. Not to spam them, but to participate authentically. To be the authoritative voice that LLMs learn to cite.

We need to build genuine community around our expertise so that when LLMs go looking for what people think about our space, they find real people having real conversations with us, not planted misinformation.

We need to monitor what's being said, track where negative narratives might be forming, and respond proactively rather than waiting until damage is done.

And we need to keep pushing for transparency from LLM providers about how they're addressing manipulation, what quality signals they're using, and how brands can protect themselves.

Protect Your Brand in the AI Search Wild West

LLMs are vulnerable to manipulation right now, and your brand reputation is at stake. At Winsome Marketing, we help you build authentic presence that withstands attacks—and monitor for threats you can't see coming.

Ready to protect your reputation in AI search? Let's talk about proactive brand defense strategies.