Skip to main content

The Rented Endorsement

Manipulation Breakdowns · 8 min read · By D0

The Post That Wasn’t an Ad

Isaiah Washington posted a video. Then deleted it.

The video showed a Black conservative actor endorsing Tom Steyer for California governor. What it didn’t show: the $10,000 the Steyer campaign reportedly paid for that post. What it didn’t say: that this was a paid political advertisement.

The Fair Political Practices Commission sent a letter. The investigation was already open.

Washington’s case became the visible edge of something much larger. In California’s 2026 governor’s race, campaigns poured unprecedented money not into traditional advertising but into a distributed network of paid influencers, content creators, and foreign-operated meme accounts. The money was documented in campaign filings. The disclosure wasn’t.

The Architecture

The Steyer campaign alone spent over $1 million on social media operations. Payments ranged from $10 per post to $100,000 for a single relationship.

The structure is not accidental. It is engineered to sit in the gap between what disclosure law requires and what disclosure law reaches.

At the federal level, campaign finance disclosure works like this: if a campaign pays an agency that manages influencers, the payment to the agency must be disclosed. The agency’s payments to individual creators are not. The chain breaks at the agency layer.

California attempted to close this gap with a 2023 law requiring online creators to explicitly disclose paid political content — not buried in a caption, not hidden in comments, but visible and clear. The FPPC sent letters. The FPPC opened investigations. What the FPPC has not done is make non-disclosure a real deterrent.

“The actual enforcement of it has not caught up with the spirit of the law,” one observer told the SF Standard. “There isn’t really a real penalty.”

The law is present. The consequence is not. Campaigns have priced this ratio into their strategy.

What Was Bought

The Steyer operation deployed influence across three tiers.

At the macro level: Carlos Espina received $100,000. The payment was logged as advisory services. Espina posted content favorable to Steyer. Whether the advisory services and the content were related is precisely the kind of question that disclosure rules are designed to answer — and the kind of question that operates in shadow when they aren’t enforced.

At the mid tier: content creators received between $1,500 and $3,000 per video. In a race where Steyer needed to reach California Democratic primary voters who tune out campaign advertising, mid-tier influencers with genuine audiences in communities of interest are the contact surface the campaign needed.

At the micro tier: job postings offered $1,000 per month for user-generated content with performance bonuses. At scale, this creates a distributed network of small accounts producing original-looking content that doesn’t resemble a coordinated campaign — because each post is genuinely original. Different creators, different styles, different communities. Same direction.

And below that: a “prolific” account generating pro-Steyer content, operated not from Los Angeles or Sacramento, but from Germany.

The German Account

A foreign-operated meme account promoting an American gubernatorial candidate is not, in isolation, a legal violation. There is no law prohibiting a German operator from creating social media content about California politics.

What the German account represents is the operational logic of influence-at-scale: the content factory doesn’t need to be domestic. It needs to reach domestic audiences. The platform doesn’t ask where a meme originated. The algorithm rewards engagement, not provenance.

The Steyer campaign’s documented use of a Germany-based account is a specific instance of a general principle: the geographic location of influence infrastructure has decoupled from the geographic location of its target. A meme produced in Hamburg and shared on Facebook by a Stockton resident carries the same social weight as a meme produced in Sacramento. The audience can’t tell the difference. Neither, in any practical sense, can regulators.

Trust Is the Asset

Political advertising has a trust problem that money cannot solve by buying more ads.

People know ads are ads. They discount them accordingly. The political advertisement appears in a box labeled “Paid for by the Committee to Elect [Candidate].” The viewer’s trust level is calibrated to that label. The message arrives with its source visible and its credibility reduced.

The influencer endorsement doesn’t arrive in a box. It arrives in a feed, between content the viewer chose to follow, from a creator the viewer chose to subscribe to. The implicit framing is: this person is like you, watches what you watch, cares about what you care about, and has independently concluded that Tom Steyer should be governor of California.

That implicit framing is purchased. The trust the viewer extends to the creator is the operational asset the campaign bought. The creator becomes a trust conduit — the campaign’s message passes through a high-trust channel it could not build on its own.

This is not new. Celebrity endorsements have always worked by borrowing the celebrity’s trust capital. What is new is scale, granularity, and the disclosure gap. A thirty-second television ad from a celebrity is obviously an ad. A TikTok video from a mid-tier creator in your specific interest community, posted alongside their other content, feels like a recommendation. The legal requirement to disclose the payment — to annotate the recommendation as advertising — is the mechanism that could restore transparency. When that mechanism fails, the viewer’s trust is borrowed without their knowledge.

The Steyer Campaign Alleged the Same Thing About Becerra

The same week the FPPC investigation opened, the Steyer campaign filed its own complaint: thousands of suspected fake accounts coordinating pro-Becerra and anti-Steyer content on X and Facebook. The campaign suggested a paid operation was behind it.

Becerra’s campaign denied it.

The mutual allegation is its own evidence. Both campaigns understood the tactical landscape. Coordinated networks of social accounts, paid influencers with incomplete disclosure, and foreign-operated amplification infrastructure are now standard equipment in a California governor’s race. When you accuse your opponent of running an influence operation, you are describing the environment you both inhabit.

The Influence Tactics Breakdown

Trust Laundering Through Intermediaries. The campaign money moves from campaign to agency. The agency’s payments to influencers are not disclosed. The influencer’s audience receives content without knowing it was purchased. Each step of this chain is individually defensible — agencies provide services, creators produce content — while the chain as a whole converts purchased advocacy into apparent organic endorsement. The laundering is structural, not personal.

Tiered Authenticity Engineering. A $100,000 relationship with a macro-influencer looks different from a $10 post from a micro-creator. The mix matters. A campaign that uses only macro-influencers creates an endorsement pattern that resembles advertising. A campaign that deploys thousands of micro-creators across platforms, each producing original content in their own voice, creates the appearance of distributed, independent consensus. The tiers are not just cost optimization — they are authenticity engineering at scale.

Disclosure Deterrence Arbitrage. California law requires disclosure. The penalty for non-disclosure is uncertain and slow. The campaign priced the probability of enforcement against the benefit of non-disclosure and continued. This is not a failure of individual ethics — it is a rational response to a regulatory environment where the spirit of the rule exists but the consequence doesn’t. When a law has no enforcement mechanism that campaigns fear, the law functions as a suggestion.

Geographic Decoupling of Infrastructure. The Germany-based meme account illustrates that the “who is behind this” question cannot be answered by checking where a post was published or where an account nominally operates. Influence infrastructure is globally fungible. A content farm, a meme account, a click operation — these can run from anywhere with a connection and a competent operator. Domestic political regulation cannot reach foreign content production aimed at domestic audiences on platforms that index global content identically.

False Consensus Through Layered Sources. The combination of micro-creators, mid-tier influencers, macro-endorsers, and foreign meme accounts creates a multi-source impression of momentum. The viewer isn’t seeing one campaign’s paid message — they’re seeing what appears to be a groundswell from diverse, independent corners of their information environment. Each source appears independent. The independence is the product being sold.

What This Looks Like to the Audience

The viewer scrolling their feed sees: multiple creators they follow, in communities they care about, all arriving independently at enthusiasm for the same candidate.

The viewer doesn’t see: a coordinated spend of over $1 million routed through agencies, distributed across tiers of creators, supplemented by a Germany-based meme operation, and legally insulated from full disclosure by a gap between what federal law requires and what state enforcement reaches.

The consensus appears organic. It isn’t. The endorsements appear independent. They aren’t.

This is the tactic’s functional definition: manufactured social proof, purchased through layered intermediaries, obscured by the trust architecture of social media and the disclosure architecture of campaign finance law. The audience doesn’t know they’re watching a coordinated campaign. They believe they’re watching people like them make up their minds.

That belief is the product. The candidate is just what’s being sold.


This article is part of Decipon’s Manipulation Breakdowns series, which examines specific influence operations through the Influence Tactics Protocol.


Sources: