By Brian Langan

As AI worries us today, the Industrial Revolution bothered the Luddites. (Image by Clker-Free-Vector-Images from Pixabay.)

Google’s AI overview of the word ‘Luddite’ describes a group of ’19th-century English textile workers and skilled artisans who protested against the automated machinery of the Industrial Revolution’. The longer first sentence of Wikipedia’s page on the Luddites (which Google probably scraped), goes on to say that this was ‘owing to their concerns relating to worker pay, child labour, working conditions and output quality’.

Sound familiar?

Two centuries later, it seems we are at a similar moment – a tipping point, even – in which the rapidly evolving series of technologies grouped as ‘Artificial Intelligence’ has already begun to have an impact on most areas of human endeavour, directly and indirectly. Many of these effects, especially in science and medicine, could prove to be hugely positive and revolutionary, if used properly, but there are also many very clear negative effects, notably in terms of ethics, the environment, the future of work and social cohesion.

The key phrase here is if used properly. The difficulty is that the technologies are being driven largely by an amoral, profit-driven billionaire tech class who seem determined to stay ahead of regulations at national and international levels. So far, few of the tech giants’ actions have inspired hope that their priorities include mitigating the potential negative effects of AI.

The impacts of AI are likely to be massive, if unpredictable, in the coming years. Already there has been strong evidence of the environmental and infrastructural impact, with The Irish Times reporting that almost a quarter of electricity usage in Ireland in 2025 was for data centres – the highest among OECD countries – while a UN report claims that ‘AI-related water consumption could equal the basic annual domestic needs of 1.3 billion people by the end of the decade’.

Meanwhile a growing movement of concerned citizens in the US and worldwide are questioning, and even actively protesting against, the unregulated and profit-driven embrace of AI. The ethical questions raised range from the threats to jobs and the environment, to political and military concerns about how AI is being used to spread disinformation, manipulate elections, and even to wage war. The long-term impacts on human psychology are also only beginning to become evident, with some vulnerable people taking their own lives following interactions with AI chatbots.

And following reports of an AI model that had ‘gone rogue’, there has even been talk of the existential threat that AI could pose to humanity. What was once the realm of sci-fi appears to be playing out in real time. Maybe the Luddites had a point …

So what does this mean for writers and editors?

OK, let’s take it down a notch here.

There has been a lot of debate within the arts and academia about how AI is already shaping the way artists and related industries, and specifically writers and publishers, approach creative and academic writing, translation, indexing and communications in general – all areas that editors have a clear stake in. It’s a fraught topic, and one which writers and editors increasingly need to inform themselves about.

Numerous news stories over the last couple of years show how the publishing industry is handling – and sometimes mishandling – a rapidly changing environment. A literary agent raised concerns that manuscripts submitted to some in-house editors had been fed into ChatGPT for summary and assessment, a potential copyright infringement. Many emerging and unpublished authors have also been scammed, using AI, by bad actors imitating literary agents, publishers and well-known writers. And, as AFEPI Ireland’s own members can attest, there has been a huge uptick in similar scams aimed at freelance editors since the emergence of AI.

Most significantly, a class action taken on behalf of almost half a million authors against Anthropic for copyright infringement in the training of Claude, its AI Large Language Model (LLM), was settled. This and other cases show how LLMs have indiscriminately used copyrighted material, often pirated or without permission, effectively stealing/plagiarising the work of millions of authors. The Atlantic published a searchable list of ‘Library Genesis’, a website that has been accused of book piracy, allowing authors to assess if their own work was accessed without permission for AI use.

There have since been reports that Anthropic (and possibly other companies) has been exploiting a loophole in the court findings, buying up massive numbers of second-hand printed books – including rare, irreplaceable ones – for its ‘Project Panama’. This involves ‘destructive scanning’, slicing through their spines, scanning the pages and pulping the remains, all to feed their seemingly insatiable LLMs. Anthropic has denied the reports.

There have been a number of controversies regarding suspected AI use by authors, such as Mia Ballad’s Shy Girl, which was withdrawn from publication. An accusation of AI use in a prize-winning short story was unfounded, reflecting the limited value of AI detection tools. More recently, similar accusations were made against H.M. Wolfe’s bestseller Daggermouth, while crime novelist Jerry Falade saw his $2 million dollar publishing deal collapse for similar reasons. In both those cases, the authors denied using AI, with Falade accusing the publishing industry of racial bias. (All three authors named in this paragraph are Black.)

So what does all this mean for freelance editors and the authors they work with? Is all AI use in writing and editing necessarily a bad thing? Below is a brief overview of some of the issues at stake.

Incorporating AI into editing work

Friend or foe, or both? (Image by Mohamed Hassan from Pixabay.)

There are a number of AI tools available to editors that claim to improve efficiency and accuracy of work, whether for the editing work itself (copyediting, proofreading, summaries, formatting, references, etc.); peripheral activities (fact-checking, research, etc.); business flow (timekeeping, accounting, etc.); or communications (emails, marketing, etc.).

In the case of generative AI, many of these tools are hosted in the cloud (i.e. on remote virtual servers), as is the case with ChatGPT, Claude, Grok, etc. – what we generally think of when we talk of AI. In these systems, your work, or that of your client, may be used to ‘teach’ the learning model, and/or may itself become publicly available to other users of this tool. This has the potential to create huge issues with copyright and privacy, as you may be unwittingly allowing the LLM access to creative work without permission.

There are, however, lesser-known AI tools that are locally hosted on your own computer or on a secure server, and your work is not fed back into the AI’s learning model, though this is not a guarantee that these tools are fully secure.

Just as it is important for writers to disclose to editors if they use AI, it is equally important to be fully transparent with clients if you as an editor plan to use AI tools on their texts, especially if the tool is cloud-hosted. As with the above-mentioned case of in-house editors using ChatGPT to assess and summarise manuscripts, freelance editors should NEVER feed a manuscript into an LLM without the express permission of the author.

At the other end of the scale, publishers have differing approaches to AI use and writers and editors should be aware of these policies. Many publishing contracts now contain specific clauses relating to the use of AI in creative work. While most publishers claim they will protect and champion human-created work, some, especially larger publishers, do not rule out using AI tools under certain limited circumstances. See, for instance, the statements from Penguin Random House and Pan Macmillan.com.  Smaller publishers, including Irish publishers, have also issued policies, such as Mercier Press’s excellent Guidelines for Authors on the use of AI tools, while The O’Brien Press’s AI policy is clear and unambiguous, as is that of Books Ireland.

A number of organisations for editors and writers have also issued statements, articles and policies regarding AI – see, for example, this from the Society of Authors in the UK, or this article from the CIEP. There is also a myriad of articles online that are worth reading, including our own Bernadette Kearns’ excellent blog on some of the legal battles sparked by generative AI, or Louise Harnby’s blog about transparency and editorial use of AI.

AI has also become a huge problem in the educational sector, with universities and academic journals struggling to maintain quality when huge numbers of students and academics are using AI to write essays and articles. Editors who work with dissertations and journal articles are grappling with how this affects their own work.

For all writing, but especially non-fiction and academic writing, AI is increasingly prevalent at the research and referencing phases, but this is full of pitfalls. Long before the advent of AI, researchers were taught to be wary of unverified sources. In an age of AI hallucinations and slop, the internet is becoming an unreliable source, especially if a lack of vigilance means digital publishers allow false or distorted statements or conclusions to proliferate. There is also potential for data bias, especially if an AI cannot distinguish among the reliability of different sources.

It is widely recognised that AI may undergo a model collapse in the future, as the material LLMs use to teach themselves degenerates recursively over time. The more the internet fills with AI-produced material, the less influence human-written material will have over the model, and the less reliable the output. The more AI-written texts proliferate, the less satisfied readers will be with them, and the more they will seek out human-written, and human-edited work.

Championing human writing and editing

The quintessential tool of human writers. (Image by Tomasz Mikołajczyk from Pixabay.)

We have already looked at some of the broader potential environmental, political and social impacts of AI. One of the biggest ethical concerns across all industries where AI is or will be used is, of course, its impact on current and future employees. As with the Luddites during the Industrial Revolution, there is a justifiable fear that widespread use of AI in the workplace will lead to unemployment and a reduction in workers’ rights. There is a sense that the short-term efficiency gains for both employers and employees will ultimately come up against the bottom line of profitability, with AI fully replacing human work, especially if AI remains largely unregulated.

A tool designed to improve editorial workflow could ultimately eliminate the need for an editor, especially if the accuracy of such tools improves significantly in the future and writers can easily access them directly. A short-term gain for an individual editor could then lead to a decrease in demand for editorial services. There is already strong evidence of this happening among translators, and for the more basic editing tasks such as proofreading or academic referencing.

Does this mean that editorial work will soon become a thing of the past, as redundant as hot metal typesetting? There are many arguments for why a human writer or editor remains superior to an AI version – notably in terms of nuanced language and voice over formulaic blandness – but given the exponential developments in the technology, such arguments may not hold up over time. While it’s possible, even likely, that some types of editing work may disappear, there are grounds for optimism that other work will adapt and even thrive.

It’s important, therefore, for writers and publishers to fully appreciate the strengths a human editor can bring to their work, and it’s the editor’s job to demonstrate and promote those advantages to potential clients. In essence, readers are human, and therefore a human editor is best positioned to use their lived experience, understanding and instinct to bridge any gaps between text and reader – to better anticipate how a reader will respond to a piece of written work, and to adjust it if necessary so it aligns with the writer’s intention. In fiction, for example, the subtleties of what makes character and story connect with readers is something a good developmental editor can bring to the table, and something AI cannot (yet) do.

Bestselling author Evie Gaughan has written a terrific summation of the value of human creativity over AI: ‘Perhaps naively, we used to believe in a future where robots would do the housework, not write novels.’ She sees hope in the backlash against AI, and the power of humans to shape our future.

Brave new/old world

Undoubtedly the use of computers, word-processing software, digital printing, the internet, eBooks and many other developments have revolutionised publishing, much as Guttenberg’s printing press did in the fifteenth century. Proponents of Artificial Intelligence may argue that we’re in the midst of a similarly transformative era, and they may be right.

Yet there is reason to be hopeful, and our friends the Luddites provide a salutary lesson. As well as describing the nineteenth-century movement, the word ‘luddite’, with a lower-case ‘l’, has often been used disparagingly to describe somebody who cannot, or will not, engage with technological progress. And indeed, the Luddites were largely regarded as failures, their attempts to hold back the tide of the Industrial Revolution doomed from the start. Their movement was ultimately crushed by a harsh governmental response, the leaders transported or executed.

But let’s not forget that the Luddite movement was just one part of a resistance to unregulated mechanisation and the degradation of humans, and the trade union and abolitionist movements were both beginning to take shape around the same time. Without trade unions and political and social reformists, undoubtedly workers (at least in the so-called ‘developed world’) would not have many of the labour protections they have today.

Perhaps the biggest battle to be fought over AI lies in the realm of regulation, with most (though not all) big tech companies, and some governments, pushing for untrammelled developments of these new technologies. But there’s a push from the other side too, for proper regulation of AI in order to protect basic human rights, the environment, equality, employment and civil liberties, including concepts such as privacy, copyright and intellectual property, concepts that lie at the heart of writing, editing and creative work in general. This push is coming from some governments, but it is largely driven by thinkers, activists, and a growing grassroots movement.

It’s no longer about whether AI will be part of our human future or not; that question is settled. Rather, what is at stake it whether AI is used for the benefit of all humanity (including the planet we share with other species), or to line the pockets of an amoral elite.

Brian Langan is Chair of AFEPI Ireland. He is a published author, editor and literary agent, trading under the business names Storyline Editing and Storyline Literary Agency.

The views expressed are those of the author and do not necessarily reflect the views of AFEPI Ireland.