US-based journalist response platform Qwoted has banned hundreds of PR agencies, executives and ‘experts’ who either appear not to exist, use AI to write or opine on areas outside their expertise.
CEO of Qwoted Dan Simon said fake AI-generated content is booming and weeding it out is complicated.
“Zooming out for a second, you’ve got real agencies representing fake experts with fake expertise. You’ve got real agencies with fake experts and real agencies with real experts with fake expertise, and fake agencies with fake experts with fake expertise.
“AI has just poured petrol on the disinformation fire.”
One profile banned by Qwoted is James Allen of Billpin, a prolific finance expert featured dozens of times in top finance publications from CBS News to the New York Post to Business Insider (see main picture).
Qwoted flagged his account for repeatedly submitting AI-generated content (the site uses AI detector Pangram to highlight potential abuse), and for failing to pass verification processes.
Allen is a curiously hard individual to pin down, for such a successful entrepreneur: America’s Financial Educators Council also raised questions over Allen’s claimed qualifications.
Since that point, Allen’s Linkedin profile has disappeared, and he has not replied to Press Gazette’s multiple requests to verify he exists.
The site Billpin.com is now blank.
Fake experts are mainly deployed as a dubious SEO tactic, driving links from reputable newsbrands to sites that want to improve their visibility in Google search results.
Simon said his team initially did not understand why organisations were creating fake experts – but there are actually several reasons.
Facets of fakery
Qwoted head of user experience Shelby Bridges said there are various degrees of fake response: “One is just misrepresentation. I actually do own the company. I run this business, but I hand my account over to an executive assistant that works for me, or an EA that’s in Nigeria.
“Then that person, I just tell them all day long: ‘Use this platform [to respond to journalist queries]. I don’t have time because I’m running a business, I don’t have time to get you content or anything else.’
“That person is just trying to spam content out, they don’t care if the person is just going in and trying to spin up as much content, trying to spam content out. They don’t care what they’re pitching. They don’t know the topics that they’re pitching on. They’re just taking questions that they find anywhere on the internet from journalists, and they are answering them through ChatGPT.”
Bridges, who started working for the company in 2022, said that the amount of spammy content began to expand rapidly from November of that year, with the public release of ChatGPT, and has since become a “game of cat and mouse” for the company.
Bridges said ‘experts’ who answer multiple queries on any subject hoping to land links in reputable publications are the easiest to spot, because it tends to be a person with a very specific frame of expertise who is suddenly commenting on wildly off-topic things.
Take Grace Carter of UKWritings.com (also pictured above), who has published work online on a range of topics ranging from advertising technology to Big Data to pet sitting to writing UX proposals to writing a CV as an architect.
In May 2025, she even offered a first-person account of living with germaphobia for Stylist magazine – which is remarkable, given that it seems highly likely she does not exist.
The company UK Writings was highlighted as problematic by Qwoted, and appears to be based in Bangladesh.
Email queries to both ‘Grace’ and UK Writings requesting verification of her identity went unanswered.
AI-generated Dan Brown offers publishers quotes on anything
Dan Brown of London-based Handy Cleaners, a company that offers end-of-tenancy cleaning services and home and office cleaning, also has surprisingly wide-ranging expertise.
Brown has commented on everything from human resources to professional car cleaning to how to get odours out of gym clothes to surprising things one can clean in a dishwasher.
His expertise appears more than 40 times in mainstream publications from Homes and Gardens to Living Etc.
Brown’s car-cleaning tips come up as 100% AI-generated according to AI detector Pangram, and his profile picture on his curiously empty Linkedin profile rates as 9/10 artificially generated on Identifai.
Press Gazette left both phone and email messages with Handy Cleaners asking to speak to Brown, and received no reply, and the email address he used to correspond with Qwoted bounced.
A request via his Linkedin profile, which lists no previous employment or education, went unanswered.
Qwoted uses both a human vetting team and technology that flags accounts for, for example, pitching a large number of times per day.
Bridges said the network deals with upwards of 15 violations per day, ranging from outright fake personas to AI-generated content, and 50 clients are banned from signing up per day.
Bridges said: “You can immediately look at that content and just realise what it is. There’s not a lot of content actually there. When you look at it, it’s just words. ChatGPT is an echo chamber.”
Bridges said there are also “spammy offshore agencies”, which often fabricate content, either using outright fake people or real people whose images are attached to low-quality content, and tend to be based in Philippines, Pakistan, India or Ukraine.
Journalists are not always bothered about fake content
Simon said the reactions to Qwoted’s efforts to stop AI-generated spam can be surprising: “I’ve had reporters say this to me, ‘I don’t mind if this expert, who is a real expert, uses GPT over their own commentary,’ which is a weird thing. We never expected to hear it.”
He said: “Today we stand very much on the side of more human content, less slop. We are playing whack-a-mole with it, essentially, and trying to figure out where to draw the line on behalf of the media.”
“We believe that public social channels like X and Facebook, and especially reporters’ email inboxes, are very unhealthy spaces right now because Alphabet, Musk, Meta and their ilk have no incentive to keep them safe.”
Qwoted trialled a tool for detecting AI content, available to both PR and journalist users of the site, but has since limited its use on the pitching side.
Simon said: “Guess who uses it? Not reporters: it’s bad actors who want to keep using it and then keep tweaking it until they get a 100% human score.”
Email pged@pressgazette.co.uk to point out mistakes, provide story tips or send in a letter for publication on our "Letters Page" blog
