'Carelessness rife' in UK publishing fact-checking as concerns grow over workloads, AI and 'status bias'

“At the publishing house I work for now, fact-checking is not something we do as standard. It isn’t part of a normal schedule for a non-fiction book. I’ve worked in the team for nearly a decade and I’ve probably done fact-checking on 10 different books: but that’s probably a high estimate.”

A freelance copy editor, who spoke to The Bookseller anonymously, said: "Even at a distance [from in-house] I can confirm that, yes, most editors working in publishing are dealing with a very heavy workload and are visibly stressed. It’s easy to see how seemingly small details… can slip through the net. They just don’t have the time, and so many of these books are being churned out on extremely tight schedules."

The concept of "status bias" was also cited as making it more difficult to challenge expert authors. Writer Pragya Agarwal said: “Short timeline for publishing with some high-profile books set to be released before they are perhaps ready, less budget for fact-checking, and so perhaps there isn’t as rigorous fact-checking with well-known names as there should be because of status bias.”

She added: “When I have written about research from a discipline that I am not an expert in, I start from the basics, and I speak to people in that discipline to make sure that I have interpreted it correctly. This takes time and energy, both often in short supply when writers are under pressure, and we also know how precarious a publishing model can be in terms of advances [many authors reported major financial difficulties in a recent Bookseller survey]. In all of my books, my publishers have also been involved in fact-checking and they have caught any errors I might have made accidentally.”

One anonymous editor speaking to The Bookseller suggested corporate publishers do not want to challenge well-respected authors. “My first role in publishing was as an editor for a small literary press, where I was both text-editing and publishing books. About a third of the books we published were scholarly works, and in these situations, where I didn’t have the capacity – or the budget – to fully fact-check, my approach was to spot-check roughly 10% of facts, usually the ones that triggered my b******t radar. I think a significant part of a publisher’s job is honing their b******t radar. If I found errors, I would begin to step up the scale of the checks, and if there were enough errors to concern me, we would ask a peer academic to review the text. I then moved to a [larger] UK publisher and was shocked by the relaxed approach to fact-checking. Because authors were often esteemed academics or well-respected journalists, we were expected to trust their work totally.”

Another editor, who spoke to The Bookseller on the condition of anonymity, echoed this: "There’s not a point in the publishing process when we say, ’OK, now it’s time for the fact-checking’. We work on the basis that the author is using reliable sources and doing their own fact-checking. If something looks or feels off, I’ll ask for more information or clarification but, ultimately, we work on trust and the burden of responsibility lies with the author."

Another editor, also working for a corporate publisher, said: “In one book, I recall an author stated a disproven theory about the formation of the moon. I queried it, but the author – whose feelings were very sensitive throughout the edit – wanted to retain it, and it was ultimately accepted among other ‘facts’ in order to placate the author. The relationship between the agent, author and publisher trumped the integrity of the information. The book went on to win an award.

“There were other books I worked on – ultimately successful, popular books – in which studies were interpreted by journalists in ways that may not have been mathematically correct. There was no process for fact-checking at this press other than relying on freelance editors to independently query points: a lot of responsibility to give someone who we in the publishing team never meet. Thankfully, for some books that were being co-published in the US by academic presses, the US publishers would commission peer reviews and those corrections were incorporated into the edit.”

Author and former publisher Isabel Thomas echoed this: “As you’re no doubt finding, checking processes vary hugely between publishers, and sadly there are many examples of style over inaccurate substance creeping through and even winning awards.”

Another editor said: “In my current role as a commercial publisher, we occasionally, though not consistently, will commission a peer review – not always paid; often by a friend of the author’s. Most of our books are memoir, and so facts are legal matters that we take very seriously with internal and external legal reads. We also encourage the authors to keep a record as they write of all legal background and the changes they make to disguise identities/events.”    

Writer Dhruti Shah warned that the lack of fact-checking in books endangers historical record. “Books are cited in scholarly papers, the press and educational fields. If those facts in the book are wrong in the first place – and there’s no guarantee there will be a reprint – there’s still the risk it will get repeated, until eventually it becomes the believed information and the dominant narrative and that’s when we are in dangerous territory.”

Shah added: “When people talk about disinformation and misinformation – so my bread and butter having worked in verification for a decade – they talk about news coverage, social media, but there needs to be more conversation about the role books and publishers play in cementing wrong information as fact.”

One non-fiction writer told The Bookseller they had to organise and pay for their own sensitivity reader. “I was writing about topics including transwomen, queer and black feminism – they [the publisher] did not care,” they said.

Writer and editor Simon Guerrier, who has collaborated with the Writers’ Guild of Great Britain on this issue, spoke to The Bookseller in a personal capacity. He revealed that the books committee of the Guild has been “drafting guidelines for writers of non-fiction, in response to the same things cropping up repeatedly in enquiries from members”.

He added: “There are things I get asked by new and would-be writers, and things I wish I’d known when I was starting out… there are questions about how you gather material, the rights that an interviewee or archive holder might have over what you write, and about your copyright as the author.”

Like many, he believes AI further complicates the fact-checking process, adding: “My sense from what I’ve seen so far is that AI has far more difficulty in identifying credible sources in the same way.”

Agarwal agreed: “I also know that some writers are using AI as a shortcut to research, and while there are some advantages to this approach, again this will bring in errors because AI is not free of biases.” 

Another editor, who spoke anonymously, said: “I think AI will change things in the future if it improves and gets better at linking users to sources of information, and could potentially save a great deal of research time. At the moment, I don’t think it is useful or reliable as a way to check facts.

“I think when mistakes are made, it is unfair to lay all the blame on the author or one individual editor, or to judge them too harshly for an innocent mistake – if it is publicly rectified and the book is reprinted. Manuscripts go through many hands before they are finalised for printing and, like typos, some errors are simply missed.

“Everyone working in non-fiction publishing, including the authors, are on tight schedules and grappling with huge amounts of information condensed into a few hundred pages. PhD students have four-plus years to produce one thesis of a similar length to many non-fiction books produced in a fraction of that time. Authors and editors are working as hard as they can in a high-pressure industry.”      

American author Naomi Klein, who won last year’s Women’s Prize for Non-Fiction, told The Bookseller: “I fact-check my own books, but it’s a privilege. I’m one of a small group of writers who are paid advances large enough to be able to do this, [and] it’s been the case since No Logo. On No Logo, I didn’t have a fact-checker, I had a little bit of research assistance but, since then, I can afford magazine-style fact-checking for all my books. But we do need in-house fact-checking and not just lawyering – that’s about not getting sued – fact-checking is about accuracy. I was lucky with this book (Doppelganger, Penguin) that my copy editor also fact-checked it and found some errors. I pay out of my advance and I work at a university so I hire three graduate students from there.”

The Bookseller approached a number of publishers for comment but only Cambridge University Press (CUP) and Manchester University Press (MUP) responded with comments. Tom Dark, trade non-fiction publisher at MUP, emphasised how authors must declare if they have used AI. “While the author is responsible for checking sources, especially interviews, direct quotes and oral history, the publisher also needs to conduct its own checks and offer authors guidance on how to cite and acknowledge a variety of sources.    

“This is obviously a pressing issue with the rise of Generative AI. We and many other publishers don’t, and indeed couldn’t prohibit authors from using this technology. We ask that any such use by authors is clearly flagged to us, and our traditional expectations still hold in terms of accuracy and rigour. As a University Press, these elements are crucial to all that we do.

“We still operate a blind peer review process so that every book, including every trade book, is read by external experts at the proposal and manuscript stages. Along with development editing, this offers an additional layer of fact-checking and quality assurance.”

A CUP spokesperson said: “The integrity of our academic content and publishing process is paramount and we have robust publishing ethics, principles and guidelines in place to ensure this. The peer-review process plays a critical role in the evaluation and development of scholarly material submitted for publication, and we have dedicated teams and rigorous editorial processes which includes copy-editing, typesetting and proof-reading before final publication.”