New ways that AI threatens the livelihoods of journalists seem to emerge every week. Press Gazette is compiling a growing list of AI-related journalistic cock-ups, from stories containing Chat-GPT “hallucinations”, to ones that quote fake AI experts.
Now there’s a new problem. Several freelance journalists have recently found themselves being impersonated by someone sending pitches to editors and requests to scientists for interviews under their name, with alarmingly plausible story suggestions, which seem to be AI-generated.
At least one of the pitches was commissioned and after heavy editing made it onto the online version of Science, one of the world’s top peer-reviewed science journals. It has since been taken down after the impersonation was discovered. “It’s safe to say from having seen the emails between the editors that there was damage to my reputation in the process,” says the victim, Caroline Williams, a British freelancer and author.
Caroline, who is my friend, rang alarm bells about the new phenomenon when she posted on LinkedIn that she had discovered fake pitches were being sent to editors and one of her good contacts.
She was first contacted by an editor at Nature who had become suspicious after an email exchange with “fake Caroline”. The editor sent the email chain to Caroline, who shared them with me. To me, they have a whiff of AI about them, but are fairly convincing.
The writer is clearly pretending to be the real Caroline, rather than just being someone who shares her name, as they provide links to real Caroline’s previous articles.
The topic of both the Science piece and the proposed Nature article are bang in Caroline’s area of expertise as she has recently published a book on the subject – interoception. This is the way we unconsciously and sometimes consciously sense our body’s inner workings, such as our heartbeat.
Alarmingly, the pitch to Nature was for an article that rebutted the main premise of Caroline’s book and her previous articles. “The story proposed is trying to say that interoception research is flawed and its impact is overblown,” says Caroline. “So it almost feels like “she” is using my name to discredit my work while also discrediting my work by writing a badly-worded pitch and nonsensical follow-up emails.
“As a freelancer you live or die by your reputation, and it doesn’t take many bad pitches for editors to stop taking you seriously.”
Linda Geddes, another British freelancer friend of mine, was the target of two fake pitches that were eventually spotted by the same editor at Nature. One was about Circadian rhythms, one of Linda’s specialisms, while the other was about AI drug design.
“It feels a bit like this this person’s getting through the first hurdles and then failing once editors start getting into the nitty gritty,” Linda said, although we were speaking before the Science article by fake Caroline was discovered, last week.
A US medical freelance journalist called Sara Novak is the third targeted person I know about. Her faker pitched a publication called Chemical & Engineering News for an article on slimming jabs – again, her area.
So, the faker, or fakers’ method seems clear: pick an established medical journalist, create a plausible sounding email address, use AI to pitch to reputable publications, and then use AI to write the final article.
In the case of the accepted Science article, the faker earned $1,000, although it may be unlikely that Science would use that writer again, judging by the comments they made about it during the editing process.
Still, it doesn’t seem the most lucrative way to use AI for scamming. Another possible motivation is suggested by Press Gazette’s discovery of the “fake experts” – entirely fictional people being quoted in newspaper articles, which has been going on for several years.
In that case, the purpose seems to have been PR-related, generating mentions of brand names and websites in the articles, for publicity or to boost website SEO rankings. If editors are (hopefully) becoming more suspicious of PR press releases quoting fake experts, perhaps fake journalists can get past that barrier and be another route to the sought-after exposure.
Richard Fisher, an editor at Aeon magazine wonders if it may be part of a more complicated scamming process.
A mention of supplements or other products in a feature in Nature or Science could be highly valuable – potentially worth more than any fee for the article itself.
The fake journalists might be hoping to extort bribes from supplement sellers, health influencers, and so on. “The scammer would say ‘feel free to contact my editor to confirm the commission, but let’s keep the fee between us because it’s not allowed’,” says Richard.
The mystery deepens. For now, journalists might be advised to take precautions against being faked – perhaps by regularly googling their own name, and checking on Muck Rack for articles they don’t recognise. And editors clearly need to be on their guard.
