In January this year, the UK Press Gazette published a remarkable and troubling investigation. Journalist Rob Waugh named more than fifty apparently fake experts — invented spokespeople, complete with AI-generated photographs and biographies — who had between them appeared in British newspapers, magazines and websites more than a thousand times.
The story was written primarily as a warning to journalists. But its implications reach considerably further. For lawyers conducting due diligence, wealth managers assessing counterparties, and communications professionals trying to understand what is true before they advise their clients, the fake expert problem is not a media curiosity. It is a direct threat to the quality of the intelligence on which decisions are made.
What is actually happening
The mechanism is straightforward, even if its consequences are not. Organisations — in the cases identified by Press Gazette, mostly commercial businesses seeking to improve their search engine rankings — commission the creation of fictitious experts. These individuals are given names, professional titles, photographs (generated by AI), and detailed biographies. They are then offered to journalists as commentators on topics ranging from interior design to personal finance to mental health.
Journalists, under time pressure and facing an ever-growing volume of press office content, quote them. The quotes appear in national newspapers, regional titles and online publications. The fabricated expert acquires a body of published work. The commissioning organisation gains the search engine links it was seeking. And the information environment absorbs another layer of fiction.
Waugh's investigation found one fictitious psychiatrist — a "Barbara Santini" supposedly working for an adult retailer — who had appeared in British media dozens of times before being exposed as non-existent. Three linked agencies operating under names including SignalTheNews and RelayTheUpdate had, according to Press Gazette, bombarded UK journalists with fabricated case studies including invented lottery winners whose stories ran in national and regional press.
"The organisations behind them never reply to follow-ups and move to different email addresses to avoid being blocked."
Pressflow co-founder Jelena Skene told Press Gazette that the problem intensified sharply in 2025, with her service having to reject 25 out of 37 new applicants during one surge due to concerns about their authenticity. The worst cases were identified within minutes of sign-up — automated AI systems generating and submitting fabricated commentary at scale.
Why this matters beyond journalism
The immediate damage is to journalism itself — and that damage is real. When a newspaper publishes a fabricated expert, it erodes the trust that makes authoritative reporting possible.
But for International Insight's clients, the consequences are more direct and more immediately consequential.
Consider due diligence. A standard element of pre-transaction research involves examining the public record on individuals and organisations — what they have said, where they have appeared, what their professional standing appears to be. If that public record has been deliberately seeded with fabricated commentary, fabricated credentials and fictitious corroboration, the due diligence process is compromised at source. You are not assessing the reality of a counterparty. You are assessing the fiction they — or someone acting for them — has constructed.
Consider litigation intelligence. Building a picture of an opposing party's background, credibility and behaviour frequently involves open-source research. Statements they have made, positions they have held, expertise they have claimed. If any of that can be manufactured at scale and inserted into apparently reputable publications, the evidential value of open-source material is materially reduced.
Consider reputational assessment. Communications professionals and their clients rely on understanding what the media record says about individuals and organisations. A contaminated media record — one in which AI-generated content has been published alongside genuine journalism, often indistinguishably — creates serious problems for anyone trying to separate signal from noise.
Open-source intelligence has always required verification. What has changed is the scale and sophistication of deliberate contamination. AI-generated content can now be produced, placed and indexed faster than any traditional verification process can track it.
The answer is not to abandon open-source research. It is to treat it as a starting point rather than a conclusion — and to combine it with human intelligence that cannot be faked: real sources, real conversations, real corroboration from people with direct knowledge.
That is precisely what International Insight does.
The SEO economy and its consequences
It is worth understanding the economic logic driving this problem, because it illuminates how deeply embedded it has become.
A link from a high-quality news publication to a commercial website is enormously valuable in search engine terms. It improves rankings, drives traffic and generates revenue. The value of such a link far exceeds the cost of producing an AI-generated expert, placing them with a press response service and waiting for a journalist to bite.
As Press Gazette notes, the businesses identified in its investigation range from plumbing comparison sites to retailers to ski companies. These are not fringe operators. They are mainstream commercial organisations that have — knowingly or unknowingly — commissioned the systematic fabrication of expert commentary that has then entered the published record.
The financial incentive is substantial. The reputational risk, until exposure, is minimal. The damage to the broader information environment is diffuse and hard to quantify. This is precisely the kind of market failure that regulators struggle to address and that bad actors therefore continue to exploit.
A familiar pattern
For those of us who have spent careers in investigative journalism, there is something familiar about this. The techniques are new — AI generation, automated placement, synthetic identities — but the underlying strategy is not. It is the deliberate construction of a false narrative, inserted into credible channels, designed to be mistaken for truth.
We have seen this in corporate disputes, where parties seed false information about opponents into trade publications. We have seen it in political contexts, where fabricated experts lend apparent authority to contested claims. We have seen it in litigation, where manufactured documentation is designed to shape the questions lawyers ask.
What AI has done is reduce the cost and increase the scale of deception. The intent is the same. The countermeasure is also the same: rigorous human verification, source by source, claim by claim.
"We ask real questions of real people and get answers that go to the heart of any issue. No algorithm provides that."
What to do about it
For professionals who depend on accurate information, the practical implications are clear.
First, treat the published record as a starting point, not a destination. The appearance of an expert in a reputable publication is no longer sufficient evidence of their existence, let alone their expertise. Verification of identity and credentials — genuine verification, not a Google search — has become a basic due diligence requirement.
Second, weight human-sourced intelligence more heavily. A source with direct knowledge of a situation, verified through proper investigative channels, is worth more than any volume of published commentary in a contaminated information environment. The value of real human intelligence — the kind that cannot be generated by an AI system — has increased precisely because so much of the surrounding information landscape has been degraded.
Third, be particularly alert in contested situations. Litigation, major transactions, reputational disputes — these are exactly the contexts in which motivated parties have the strongest incentive to manipulate the information available. The more consequential the decision, the more important it is to go beyond what is findable online.
Press Gazette's Rob Waugh and his colleagues are doing important work in documenting and naming this problem. The full investigation, including the live database of fake experts identified to date, is essential reading for anyone who works with media intelligence.
But naming the problem is only the beginning. The structural conditions that created it — the economic incentives, the time pressures on journalists, the scalability of AI generation — are not going away. The information environment will continue to be contaminated. The question is whether the intelligence you rely on is built to withstand that contamination.
At International Insight, our answer has always been the same: human sources, properly developed, rigorously verified. Not because we are sentimental about old methods, but because they are the only methods that cannot be industrially faked.