When the Crowd Isn’t Real

This is Part 3 in an ongoing series on how AI is being used to manipulate narratives, control public perception, and rewrite reality in real time.

You’ve seen it. A flood of comments all backing the same talking point. A hashtag suddenly trending. A politician says something wild, and within minutes, hundreds of "real" people are singing their praises.

But here’s the truth no one likes to admit: a lot of those “people” might not be people at all.

This essay breaks down three tactics used to manipulate public opinion online — astroturfing, sock puppet accounts, and bot networks — and how AI is turning each of them into a high-speed propaganda machine.

These tools don’t just shape what you see. They shape what you believe everyone else thinks.

1. Astroturfing: Fake Grassroots, Real Manipulation

Astroturfing is when a political campaign, company, or government pretends a movement is coming from ordinary people. In reality, it’s top-down theater dressed up as bottom-up democracy.

It works because humans take social cues from the crowd. If a protest or campaign seems popular, we assume it reflects public will. We’re more likely to join in — or at least not speak out.

Take what happened in the early days of COVID lockdowns. Across the U.S., protests flared up overnight. It looked like fed-up Americans were spontaneously taking to the streets. But The Washington Post revealed that the outrage was manufactured. A network of conservative groups created dozens of Facebook pages, pushed the same slogans, and choreographed the message. What looked like grassroots rebellion was a scripted campaign playing out on social media.

Something similar helped power Trump’s 2016 presidential run. Cambridge Analytica got its hands on the personal data of up to 200 million Americans. Without consent, they built psychological profiles, then used Facebook ads to target voters with custom-made fear bait. Messages were tailored down to personality type, location, and even mood. It wasn’t public opinion. It was a digital mirage, made to feel like one.

Today, AI makes that kind of manipulation even easier. Entire fake movements can be built from scratch — complete with photorealistic faces, video testimonials, and fake “viral” moments. The lie doesn’t just spread faster. It feels truer.

2. Sock Puppet Accounts: One Person, Many Masks

A sock puppet is a fake identity created to push a message while hiding who’s really behind it. Usually, it’s one person — or one group — pretending to be many. The goal is simple: make an idea look more popular or more widely accepted than it really is.

It works because we trust ideas that appear to come from different people. If a dozen strangers all seem to believe the same thing, we assume it must have some truth to it.

In 2016, Russia’s Internet Research Agency ran a sprawling sock puppet operation targeting American voters. They built fake accounts posing as Black Lives Matter organizers, conservative gun owners, feminists, and more. These accounts posted memes, picked fights, and spread conspiracies. On the surface, it looked like regular people arguing online. In reality, it was a propaganda machine in disguise.

And this isn’t just a political tactic. In 2007, John Mackey, the CEO of Whole Foods, got caught posting under a fake name — Rahodeb — on Yahoo! Finance forums. He spent years praising his own company and trashing competitors, all while pretending to be a random, enthusiastic investor.

Now AI can generate fake identities at scale. It can create people with realistic faces, unique writing styles, and full backstories. You’re not just arguing with bots anymore. You’re debating an army of fictional characters designed to steer public conversation without ever revealing who’s pulling the strings.

3. Bot Networks: Automated Influence at Scale

Bot networks are swarms of fake accounts programmed to act like real users. They post, reply, like, and share. Their job is to trick the algorithm into thinking something deserves attention — and trick you into believing it's already popular.

That early burst of fake engagement is what gives certain posts traction. Platforms like X (formerly Twitter) prioritize content that gets quick interaction. So when a network of bots swarms a post, it often gets bumped into more people’s feeds. That’s how fringe ideas go viral overnight.

In 2024, investigators uncovered a Russian AI-powered bot farm doing exactly this. These weren’t clumsy spam accounts. Each one had a realistic profile picture, a backstory, and the ability to post and comment like a real person. They didn’t just repost headlines. They joined conversations, gained followers, and played the long game. The entire setup was designed to push Kremlin-backed narratives while appearing like normal users.

Older bots were easy to spot. They had generic avatars, broken grammar, and repetitive behavior. But AI changed the game. Now they argue, adapt, and even pretend to disagree with each other to mimic real debate. They don’t just hijack conversations. They become the conversation.

Why This Matters

This isn’t about spam. It’s about how the internet is being used to manufacture consent. Lies are made to look like consensus. Truth is made to look like fringe opinion.

The goal isn’t just to make you believe something.
It’s to make you believe everyone else already does.

And none of it has to be labeled. There’s no rule that says bots, sock puppets, or fake movements need to come with a warning. You won’t get a heads-up. It will just look real.

That’s what makes AI-driven propaganda so powerful. It isolates you by flooding your feed with confidence — fake, orchestrated, and relentless. When every voice you see is saying the same thing, you start to question your own judgment. You begin to wonder if you’re the one who’s wrong.

And that’s the point. Confuse you. Wear you down. Shut you up.

Next Time You See a Viral Opinion…

Ask yourself:

• Who actually benefits if I believe this?
• Does this feel real, or rehearsed?
• Am I seeing public opinion, or a performance?