AI hasn’t just made it easier to create content. It’s made it easier to manufacture the appearance of authority. And that’s everybody’s problem.

Meet Ervin Hoskins, an American freelance journalist whose commentary about the Iran war appeared in publications including the Port St. Joe Star, a small Florida newspaper, and larger international outlets.

There’s just one problem. Ervin Hoskins doesn’t exist.

According to an October 2026 investigation by OpenAI, an influence operation originating in Iran used ChatGPT to help produce and place nearly 100 articles under seven deceptive journalist identities. The Washington Post identified at least 20 news outlets carrying the campaign’s material.

We’re not talking about AI-generated posts from obscure social media accounts or anonymous Reddit threads. We’re talking about fabricated journalists publishing articles in legitimate news outlets, complete with bylines, headshots and the credibility that comes with appearing in the press.

The journalists were fictitious. The publications were real. And the credibility was borrowed.

For marketing and PR professionals, that’s a rather alarming development. We’ve spent decades convincing clients that earned media is more credible than paid advertising. Now bad actors have figured out how to manufacture the former almost as easily as the latter.

The great credibility laundering machine

Traditionally, earned media has worked because independent editorial judgment provides credibility that advertising cannot simply buy.

But what happens when that editorial judgment fails?

In 2025, Wired, Business Insider and several other publications removed articles attributed to an apparently fictitious freelancer named Margaux Blanchard. Business Insider subsequently removed around 40 questionable first-person essays involving suspect bylines.

That same year, the Chicago Sun-Times and Philadelphia Inquirer published a syndicated summer reading list containing ten nonexistent books attributed to real authors, including Isabel Allende and Andy Weir. The writer admitted using AI without adequately checking the results. The publishers had unwittingly given imaginary products the imprimatur of genuine journalism.

In October 2026, Canadian media investigations also exposed fabricated journalist identities appearing in online publications, including AI-generated personas republishing the work of legitimate news organizations.

Meanwhile, NewsGuard had identified 3,749 AI-generated news and information websites across 16 languages by June 2026, many masquerading as legitimate publications.

And credibility is already a scarce commodity. Edelman’s 2026 Trust Barometer, covering 28 countries, found that media was distrusted in 13 of them. Trust stood at just 39% in Britain, 40% in France, 44% in the United States and 33% in Japan.

Now add industrial-scale synthetic journalism to that already combustible mixture.

When AI starts believing the publicity

Here’s where things become particularly interesting for marketers.

In our PR for Robots webinar series, we’ve been explaining how credible third-party coverage can help AI search engines discover, understand and potentially recommend brands. But what happens when the information AI finds has itself been manufactured?

A fabricated article published by a genuine news outlet may look like an independent source. If AI repeats its claims, manufactured credibility can become machine-amplified credibility.

This isn’t just theoretical. NewsGuard’s September 2025 audit found that ten leading AI chatbots repeated false claims in 35% of its targeted misinformation tests. Its January 2026 audit still found a failure rate above 28%. These weren’t measurements of ordinary search accuracy, but they exposed a serious vulnerability.

If AI can’t reliably distinguish authentic authority from manufactured authority, both brands and customers have a big problem.

What if the enemy isn’t a foreign government, but your competitor?

Now imagine these techniques in commercial hands.

A fictitious investigative journalist publishes allegations about a pharmaceutical company’s safety record. A fabricated industry analyst questions a rival’s financial stability. A network of apparently independent publications reports complaints about a hotel chain, airline or automobile manufacturer.

The allegations are invented, but the headlines look real. Search engines find them. Social media circulates them. AI systems repeat them.

And there’s fresh evidence that the machinery for this kind of manipulation is already being commercialized.

In September 2026, Anthropic exposed an operation run by LKM Company, a France-based digital advertising agency, using Claude to manufacture nearly 9,000 articles in approximately 20 languages across some 70 fabricated news websites. The operation also deployed more than 250 fake commenting accounts to amplify its stories across six continents.

The agency’s operation shifted political positions depending on the interests of paying clients. Anthropic described it as a commercial influence-as-a-service operation. Although investigators found little genuine audience engagement and no evidence that the content broke through into mainstream coverage, the capacity to manufacture apparent editorial consensus was extraordinary.

PR practitioners have seen concealed sponsorship, ghostwritten commentary and planted stories for decades. What’s changed is the potential scale.

In the past, manufacturing a hundred convincing articles required writers, editors, translators and considerable coordination. Now much of that production can be automated, adapted for different countries and distributed through apparently independent publications.

There’s no verified evidence in these cases that a commercial brand commissioned such an AI operation to sabotage a competitor. But the capabilities demonstrated raise an obvious question: what happens when somebody applies them to a corporate rivalry instead of a political conflict?

Black-hat PR could become black-hat publishing.

For years, PR agencies competed for editorial attention. Now we may also have to compete against fabricated experts, synthetic journalists and coordinated misinformation.

The commercial stakes are considerable. According to Britain’s Competition and Markets Authority, 89% of UK consumers use online reviews when researching products or services. If consumer decisions depend so heavily on third-party information, the potential damage from fabricated editorial coverage is hardly difficult to imagine.

And the risks extend far beyond fake five-star reviews. An article appearing in ten publications may look like ten independent endorsements of a claim, even when every version originated from the same invented source.

Quantity can masquerade as consensus. Repetition can masquerade as proof.

The credibility crisis is also an opportunity for PR

So, what should responsible marketers and PR professionals do?

First, recognize that monitoring a brand’s reputation now means examining not only what is being said, but who’s saying it, where it originated and whether the supposed authority is genuine.

Second, make real expertise verifiable. Credentials, sources, independent reporting and demonstrable facts become increasingly important when anybody can manufacture a convincing expert.

Third, monitor AI-generated answers alongside traditional media and search results. A false story that appears in the press today could potentially influence how a brand is described tomorrow. Correcting the original source matters more than simply issuing another denial.

Finally, strengthen relationships with reputable journalists and publications that actually verify what they publish. A placement isn’t worth much if the publication’s editorial standards are indistinguishable from an automated content farm.

This isn’t an argument against using AI to write. It’s an argument against using AI to counterfeit independence, expertise and truth.

PR needs to protect the very thing it sells

The irony is that while PR professionals are working harder than ever to establish brands as credible authorities, AI is making it easier than ever for dishonest operators to manufacture the appearance of that authority.

The profession has a vested interest in protecting editorial integrity, not simply exploiting editorial access. Otherwise, the credibility that makes earned media valuable could become increasingly meaningless.

For years, PR professionals have been trying to get their clients into the news. Now we may have to spend just as much time proving the news is real.

Sources: OpenAI — Disrupting AI-enabled False Front Operations (2026); The Washington Post — Iranian Campaign Planted Fake Articles in Real U.S. Publications (2026); Anthropic — Detecting and Countering Misuse of AI (2026); Associated Press / NPR — AI-generated content in newspaper supplements (2025); NewsGuard — AI Tracking Center and AI False Claims Monitor; Edelman — 2026 Trust Barometer; UK Competition and Markets Authority — Consumer reviews research