How AI-powered ‘ghost creator’ channels are gaming YouTube politics

A growing network of AI-powered political channels has been quietly flooding YouTube with hyper-partisan “news” videos fronted by paid actors. On the surface, these channels look like ordinary political commentators. Behind the scenes, they’re driven by automated scripts, growth hacks, and a business model that values engagement and ad revenue over truth.

What are AI ‘ghost creator’ channels?

These channels present themselves as independent political commentators delivering breaking news or sharp analysis. In reality, they are run by a single content automation operation that uses AI to generate scripts, then hires freelance actors to read them straight to camera.

Instead of real journalists or commentators, the “hosts” are paid spokespeople found on platforms like Fiverr and Backstage. They perform under fake names, often across multiple channels, reading AI-written monologues designed purely to trigger outrage and clicks.

How the scam targets political engagement

The network discovered by investigators focused heavily on attacking well-known Democratic politicians such as Alexandria Ocasio-Cortez (AOC), Bernie Sanders, Gavin Newsom, and Zora Mansoori (often mispronounced in the videos). The formula is simple: pick polarizing figures, frame them as villains, and pump out endless content criticizing them.

Interestingly, the group reportedly tested an anti-Trump persona at one point, but abandoned it when it failed to perform as well. That suggests the primary motive isn’t ideology—it’s engagement. The content leans into MAGA-style narratives because that’s what gets the most views and revenue on YouTube.

AI-written scripts, human faces

Although the hosts are real people, the scripts behind them are clearly generated or heavily assisted by large language models. The videos follow rigid templates, repeat the same phrases, and reuse identical narrative structures.

Common AI tells include:

• Stock openings like “Picture a highway in California…” or “Imagine this with me…” repeated across different channels.
• Overly dramatic lines such as “That’s not a rounding error. That’s a crater.” appearing in multiple videos.
• Formulaic storytelling about politicians’ backstories, especially around AOC’s time as a bartender, with nearly identical wording.

Because actors are just reading what they’re given, they often stumble over names, mispronounce public figures, or clearly don’t understand the topics they’re discussing. Yet the videos still feel “real” enough to pass as authentic commentary for viewers scrolling quickly.

Why use actors instead of AI avatars?

With today’s tools, the same operation could have used AI avatars and synthetic voices. But they chose human actors for a reason: real faces are harder to flag and remove, and viewers are more likely to trust them.

Platforms and audiences are getting better at spotting obviously AI-generated video characters. A human host, even one reading awkward AI text, blends in more easily with the rest of YouTube’s creator ecosystem. That makes these ghost channels more resilient and more effective at spreading misinformation.

How the network was exposed

The network came to light when an AI content tracking company noticed something unusual. Their systems, built to detect AI-generated material at scale, flagged these channels because of the heavy use of machine-written language—even though the videos themselves featured real people.

From there, investigators uncovered:

• A cluster of about a dozen interlinked channels sharing the same style, scripts, and ideological slant.
• Link trees and websites that mismatched personas, sending viewers from one fake identity to another.
• An internal dashboard with YouTube analytics that had been left publicly accessible and accidentally search-engine optimized, exposing the operation’s backend.

Further digging revealed that several of the “hosts” openly advertise their services as presenters on freelance platforms, using different real names than the ones they use on YouTube.

The role of Viralocks and its founder

At the center of this operation is a content automation company that appears to be Viralocks, founded by Caleb Chan, a former University of Chicago student who also runs a successful tech YouTube channel. Chan publicly described building systems that use AI and automation to spin up an average of two new channels per day.

In social posts, he has shared insights into how he “hacked” YouTube’s algorithm, including claims that if 26 accounts post the same hashtag within 10 minutes, the algorithm treats it as breaking news and boosts it to trending. His conclusion: “manufacturing news > reporting news.”

According to his own statements over time, his company has produced thousands of videos per week for businesses, and he has bragged about ghostwriters generating billions of views for clients, including controversial AI startups. However, in more formal comments, he has described his operation as an “independent digital media publisher” that doesn’t work for political campaigns or ideological organizations, a claim that sits awkwardly beside his public growth-hacking boasts.

Why this isn’t ‘just another misinformation problem’

Online misinformation is not new, and neither are partisan channels trying to go viral. What’s different here is the scale, speed, and low cost enabled by AI plus automation.

With AI writing most of the scripts and systems handling production and publishing, a small team can:

• Launch and grow dozens of channels in a short time.
• Test different political angles and personalities, then double down on what performs best.
• Rapidly churn out content that sometimes outperforms major news outlets like the Washington Post or the New Yorker in raw engagement.

There is effectively no quality control. Facts can be outdated, distorted, or simply wrong, and it doesn’t matter as long as the videos keep viewers watching and commenting. The incentive structure rewards outrage, not accuracy.

This emerging pattern mirrors broader concerns about AI-driven media manipulation, similar to how a fake “top-performing” AI model was used to highlight the weaknesses of traditional benchmarks in another recent AI investigation.

Why older and news-hungry audiences are at risk

The channels appear to be tuned for a specific demographic: politically engaged viewers, often older, who rely heavily on YouTube and social media for their news. These audiences are valuable because they watch longer, comment more, and are deeply invested in political narratives.

For them, a new, confident-looking host in a well-lit studio can seem like just another legitimate commentator. The fact that the person is an actor reading AI-generated copy is invisible. The result is a subtle but powerful manipulation of what these viewers see, believe, and share.

Ghost networks that vanish on demand

One of the most dangerous aspects of this model is how disposable it is. Channels can be scrubbed, renamed, or relaunched under new identities at any time. Some of the accounts identified in early reporting have already been deleted or removed.

Unlike established media brands or high-profile personalities who can be pressured or held accountable, these ghost networks can appear, influence millions, and disappear with almost no trace. That makes them ideal tools for anyone—from engagement-obsessed growth hackers to well-funded political actors—who wants impact without responsibility.

What this means for the future of political content

What we’re seeing now is likely just the early phase. As AI tools, freelance marketplaces, and cheap digital production continue to improve, it will become even easier to:

• Spin up hundreds of coordinated channels around an election or hot-button topic.
• Micro-target narratives at specific demographics or regions.
• Blend paid actors, AI scripts, and synthetic media into content that feels indistinguishable from authentic grassroots commentary.

Campaigns, super PACs, and interest groups with far more money and data than a single growth hacker could professionalize these tactics at scale. As one detailed report on this phenomenon put it, these ghost creator operations offer only a “tiny glimpse” of what’s coming—an idea explored more deeply in this breakdown of AI-generated political ghost creators on YouTube.

How to spot AI-driven ‘slopaganda’

While these channels are designed to look and feel authentic, there are patterns you can watch for:

• Recycled phrases and structures: If multiple channels use the same metaphors, openings, or oddly dramatic lines, AI scripting is likely involved.
• Hosts who seem disconnected from the topic: Mispronouncing well-known names, stumbling through key terms, or sounding like they’re reading unfamiliar text can be a red flag.
• Overly generic branding: Names like “News Reports,” “Daily Politics,” or “Talk Show” with no clear background, biography, or verifiable history.
• Thin or inconsistent online presence: Bios that don’t match across platforms, broken or mismatched link trees, and websites that feel auto-generated.
• Relentless focus on a few enemies: Dozens of videos hammering the same politicians with little nuance or original reporting.

Practical steps to protect yourself

In a media landscape increasingly saturated with AI-driven content, a few habits can make a big difference:

• Check who’s behind the channel: Look for a real, consistent identity, not just a first name and a vague bio.
• Cross-check claims: If a video makes a bold or shocking claim, verify it against multiple reputable outlets.
• Be wary of emotional manipulation: If every video is framed to make you furious or terrified, you’re probably being optimized for engagement, not informed.
• Support transparent creators: Prioritize channels and outlets that are clear about who they are, how they’re funded, and how they source information.

Why boosting authentic creators matters

One of the few effective counters to AI-powered slopaganda is to actively support creators and organizations that are transparent, accountable, and committed to accuracy. Algorithms respond to engagement. Sharing, liking, and commenting on trustworthy content helps it compete with automated junk.

This doesn’t mean blindly trusting any one source. It means building your own set of trusted voices—across the political spectrum if possible—and helping those voices reach people who might otherwise be pulled into ghost networks.

The bottom line

AI has made it easier than ever to manufacture the appearance of grassroots political commentary at scale. A small operation can now deploy dozens of fake personas, powered by AI scripts and paid actors, to shape what millions of people see when they open YouTube.

As these tactics spread, the responsibility shifts partly to viewers: to be more skeptical, to look beyond the face on the screen, and to reward creators who value truth over virality. The AI slopaganda wave is here—but recognizing it is the first step toward resisting it.