t started with an entirely ordinary interaction on LinkedIn. Someone had tagged me in a post, saying how much they'd enjoyed contributing to a New York Post article I'd written about "chronoworking". As the author of a book about circadian rhythms, it was exactly the sort of subject I might write about. Except I'd never heard of this person, let alone interviewed them.

I clicked on the link and my stomach dropped. There, above an article about people who organise their working lives around their biological clocks, was my name: Linda Geddes.

It was the latest in a string of increasingly bizarre events that had begun three weeks earlier, when a friend who works as an editor at the scientific journal Nature texted to ask whether I'd been pitching stories to one of her colleagues. I hadn't.

It turned out that someone had been approaching Nature editors using my name and an email address that wasn't mine, referencing articles I'd previously written to make their pitches more convincing. Worse, the journal had recently encountered another impersonator posing as Caroline Williams, a former colleague of mine. That person had convinced an editor at Science to commission and publish an article under Caroline's name, receiving about $1,000 for it.

Further digging revealed that Fake Linda had also pitched a story to Science, unsuccessfully. Then, last week, an editor at Nature Medicine contacted me to ask whether I'd been pitching to them and Nature Health as well.

I donned my sleuthing cap and discovered a potential link between Fake Linda and Fake Caroline. Entering my impersonator's email address into Outlook.com's account recovery system revealed a partially obscured Gmail recovery address. An editor at Science had independently discovered the same address pattern associated with Fake Caroline's email.

As someone whose livelihood depends on the reputation I've spent decades building, I find this deeply unsettling. Someone is potentially making money by passing themselves off as me, while putting my professional credibility at risk. Apart from one interview conducted by email, the quotes in the New York Post article were lifted from previously published features, without attribution. Many of Fake Linda's pitches also bear the hallmarks of AI-assisted writing. Even unsuccessful pitches could damage relationships with editors.

The New York Post has since removed the article, conducted an internal review and referred the matter to the appropriate authorities, a spokesperson said.

What I find particularly baffling is the motive. Freelance science journalism is hardly a lucrative profession, and is quite niche. My best guess is that a struggling freelancer is impersonating more established peers to boost their earnings, perhaps using AI to make the process more efficient. But how would they receive payment without revealing their identity?

Other possibilities have crossed my mind. Could it be a PR professional trying to secure coverage for a client, or someone testing how easily publishers can be fooled? Perhaps it is someone with a personal grudge. But as Caroline put it: "Why us? We're nice!"

Dr Simon Miller, the director of policy, strategy and communications at the UK fraud prevention service Cifas, is equally bemused. "Ordinarily, if you're seeking to defraud, you're seeking to do it at scale, in a relatively light-touch way, using mechanised, industrialised processes," he said. "This is tantalisingly unusual. It's almost the stuff of Agatha Christie, but with 21st-century technology."

Journalists are far from the only people finding their names attached to work they didn't produce. In April, Prof Eric Topol, of the Scripps Institute in California, discovered his name on a scientific paper he knew nothing about. Four of his five supposed co-authors were similarly unaware of the paper, while the fifth could not be identified. Within hours of Topol flagging the incident on X, the journal's website had disappeared.

Topol, who is a bestselling author, said he has discovered at least 12 books on Amazon with his name on that he didn't write, including several cookbooks. "I don't even know how to cook," he said. Particularly upsetting was when a patient told him he'd read his supposed memoir. "They said: 'I was reading it, and a lot of those things don't seem to be accurate.' That's how I found out about it."

Leslie McIntosh, a vice-president of research integrity and security at Digital Science, said someone might add a respected name to a scientific paper without permission to make the work appear more credible. Fraudsters also sometimes added respected researchers as editors to make bogus journals appear legitimate and attract publication fees. Even more nightmarish would be someone deliberately publishing misleading information under someone else's name to damage their reputation. "If their name ended up on a retracted paper, they would then have to explain that retraction," she said.

Employers are encountering yet another form of impersonation. Jamie Collier, a researcher at Google Threat Intelligence, said criminals had used stolen personal information to secure remote jobs under other people's names, sometimes without the victim's knowledge.

Generative AI is making such deception easier, "mainly by stripping away the usual tell-tale red flags that hiring managers relied on to spot fraud," Collier said. Large language models can generate convincing CVs, tailored cover letters and fabricated evidence of professional experience.

"Real-time video filters and voice-cloning tools are the next step. While live deepfakes are still relatively nascent … threat actors are already testing them to bypass quick video screenings and mask regional accents," Collier said.

The Society of Authors and the Writers' Guild of Great Britain have warned of increasingly sophisticated scams involving people impersonating literary agents and publishers, in an attempt to get writers to part with money or sign away their rights.

The Society of Authors' chief executive, Anna Ganley, said she personally received between five and 10 scam emails a day, supposedly from book festivals, literary agents, publishers and others. Some writers had lost thousands of pounds after responding to genuine-sounding offers. "It's incredibly difficult for writers, because you are often looking for opportunities to market your books," she said.

Unlike those scams, Fake Linda's activities seem like a lot of work for relatively small financial gain. But the latest technology could make such impersonations easier. Miller said: "AI is allowing criminals to create sophisticated impersonations at unprecedented speed and scale."

Or multiple impersonations. I've since identified a third science journalist, Sara Novak, with a potential link to Fake Linda and Fake Caroline. Fake Sara apparently pitched a story to Chemical and Engineering News in August, which was declined. When Real Sara shared her impersonator's email address with me, I discovered it links to the same masked recovery address.

The consequences extend beyond those being impersonated. "The risk for journals and publishers is huge," said Dr Ben Johnson, the chief editor at Nature Healthwho turned down a pitch by Fake Linda. "People trust the Nature brand as a source of evidence-based quality information. If an AI-generated story from a fake journalist is submitted then I will likely spend hours editing it from AI slop into Nature prose, including all the fake quotes from people never interviewed."

All of this raises uncomfortable questions about how you can be sure you know who you're dealing with when communicating online. When I emailed the New York Post about my impersonator, it agreed to help rectify the situation, then asked: "I see that you're emailing me from a Guardian email address. Is there any other info you can provide to verify who you are?"

After discovering that Science had inadvertently published a news story by Fake Caroline, its managing news editor, John Travis, said his team had started running all pitches and drafts through AI detectors, "though they are not perfect". "We also require Zoom calls for any new writers," he said.

The experience has also put editors at Nature on higher alert. "Sadly, for early career freelancers, this makes me less likely to commission news stories from journalists that I haven't worked with before," Johnson said. "I will also have to try to find the time to have a call with new freelancers before they write, to check they are who they say they are."

Even video calls are no guarantee of identity. Google Threat Intelligence recently documented North Korea-linked hackers using real-time deepfakes to lure executives into Zoom calls with trusted peers. "They faked video and audio glitches to get the target to download a 'troubleshooting tool' that wiped out their credentials and wallets," Collier said.

Miller advises independently checking unexpected approaches through an established email address or telephone number. McIntosh advocates several layers of verification, like the multiple identity checks at airports. "You have to have multiple ways to ensure and triangulate that you're speaking to the person you think you're speaking to," she said.

For people whose identities have been hijacked, Miller advises warning professional networks and monitoring for further impersonation attempts.

There may also be legal remedies. Rupert Earle, senior counsel at the London law firm Bates Wells, said that under UK law, impersonating someone to secure paid work could constitute fraud. Victims could also have civil claims for "passing off" – misleading people into believing that work or services come from an established person or business – or false attribution, if work they didn't create was published under their name.

However, identifying the perpetrator and taking legal action can be difficult, particularly when the publications are overseas and the impersonator's location is unknown. Despite uncovering several clues linking Fake Linda and Fake Caroline, we still don't know who is behind the accounts.

But the hunt for Fake Linda continues. I'll be reporting the incident to the UK's Report Fraud service. I've also sent a cease-and-desist email to my impersonator, demanding they stop.