Every generation seems to encounter a technology—or a wider social issue—that initially feels enormous.
There was Y2K. There was the arrival of microwave ovens in ordinary homes, accompanied by fears about safety and radiation. There was a period when acid rain was discussed so widely that even Diff’rent Strokes devoted an episode to it.
Some of those concerns were justified. Some risks were reduced through research, regulation, preparation and changing practices. Other fears became less overwhelming as the technology became familiar and people developed a clearer understanding of what it could—and could not—do.
I wonder whether artificial intelligence is another of those moments.
That does not mean the concerns surrounding AI should be dismissed. Questions about copyright, consent, ownership, employment, transparency, creative integrity and the use of existing work to train AI systems are legitimate.
Authors have every reason to care about how their intellectual property is obtained, used and protected.
We should oppose exploitation. We should expect companies to be transparent about how their systems work. Authors should also be free to decide that particular forms of AI do not belong in their own creative processes.
However, the conversation becomes much more complicated when AI is discussed as though it exists only when someone opens a generative program and asks it to produce words or pictures.
For authors, AI is already present across much of the commercial environment in which we work.
The websites we use
Many authors maintain websites through WordPress or use Elementor to build and manage their pages.
WordPress is one well-known example because it is widely used by authors, but it is certainly not the only website platform involved.
Authors also use Wix, Squarespace, Shopify, GoDaddy and a wide range of hosting providers, website builders, templates, plugins and third-party services.
Some authors build their own sites. Others employ designers who choose and manage the software on their behalf.
WordPress.com now offers AI tools that can assist with website creation, layouts, written content and images. Elementor also offers AI-assisted text, images, page structures, HTML and CSS directly within its page editor.
That does not mean every WordPress or Elementor website was created with AI.
The presence of an AI feature within a platform does not prove that every user activated it. Many authors built their websites long before these features existed, and many continue to work manually within the same platforms.
But the technology is now part of the wider website environment.
If our position is that any association with AI is unacceptable, are we prepared to pull down our websites?
Would we pay a designer to rebuild them? Which software would that designer use? Could we be certain that none of their coding, design, accessibility, security, optimisation or hosting tools incorporated AI or machine learning?
The design services we use
Authors routinely use Canva and Adobe for social-media graphics, newsletters, bookmarks, banners, advertisements, launch material and other promotional content.
Both companies now openly offer AI-assisted design features.
Again, using Canva or Adobe does not prove that an individual author generated their artwork using AI. Both services still allow people to create designs manually using templates, photographs, licensed elements and ordinary editing tools.
But if the line is not about whether we personally use an AI feature, and is instead drawn at using any company that offers AI, are we prepared to close those accounts?
Would we stop using Officeworks or VistaPrint because they accept artwork created through platforms that include AI tools?
Would we ask every designer which software they use, which version they use and whether any part of that software relies on machine learning?
The newsletters we send
Newsletters are one of the most commonly recommended tools for independent authors.
MailerLite, Mailchimp, Brevo, Kit, beehiiv and other providers have introduced AI-assisted writing, subject-line generation, campaign creation, audience analysis or send-time optimisation.
An author may never use those features.
They may write every word themselves, select every image manually and schedule every email personally.
Yet the platform may still use automated systems to detect spam, analyse delivery, assess engagement, segment subscribers or recommend when an email should be sent.
Where, then, do we draw the line?
Is the line at AI-generated newsletter content?
Is it at AI-assisted subject lines?
Is it at automated audience analysis?
Or does the entire newsletter provider become unacceptable once AI exists somewhere within its services?
The payment systems we rely upon
Suppose an author decides to avoid every design or website platform associated with AI and instead employs human professionals.
How will those professionals be paid?
Stripe, PayPal and Square all use AI or machine learning in areas such as fraud detection, transaction analysis and risk assessment.
These systems are not writing books or designing covers. They are helping payment processors decide whether a transaction appears legitimate.
Most of us would probably prefer our payment providers to use effective fraud-detection systems.
But it is still AI operating within a service we rely upon.
Does paying a human designer through Stripe, PayPal or Square make the transaction unacceptable?
Does it matter that the AI is being used for security rather than creative production?
Most people would instinctively say these are different uses.
I agree.
But once we acknowledge that difference, the ethical question can no longer be reduced to simply “AI or no AI.”
We must ask what the technology is doing, what data it relies upon, whether consent was obtained, whether anyone is harmed, and who remains responsible for the final result.
The publishing platforms we depend upon
Authors also rely on Amazon KDP, Draft2Digital, IngramSpark and other distributors and retailers.
Amazon KDP distinguishes between AI-generated and AI-assisted content. Its current guidelines require disclosure of AI-generated text, images and translations, while treating editing, refining, error-checking and brainstorming differently.
Draft2Digital also distinguishes between AI-assisted material and books generated almost entirely by machines without substantial human involvement.
That distinction matters.
Asking a program to generate an entire novel is not necessarily the same activity as using software to identify repeated words, check punctuation, test readability, resize an image or suggest categories.
Even long-established author tools now raise questions.
ProWritingAid openly describes itself as using AI-powered analysis. Other programs may rely on automation, algorithms or machine learning without users necessarily knowing which technical label applies. Oh and don’t forget CoPilot!
Then there are services such as Booklinker, Books2Read, BookVault, Publisher Rocket and Atticus.
These use automation, data analysis, link localisation, formatting, fulfilment or market-research systems. Their public information does not always make clear whether particular functions should be described as AI, machine learning or conventional software automation.
That uncertainty is itself important.
How many authors could confidently identify every algorithm, plugin, third-party provider and background process used by every company supporting their publishing business?
The systems that decide who sees us
Authors also advertise through Amazon, Meta, Google, BookBub and other online platforms.
We post on Facebook, Instagram, TikTok and YouTube.
We depend on recommendation systems, search rankings and automated content distribution to put our work in front of readers.
Those systems affect visibility.
They may help decide which advertisement is shown, which post appears in a reader’s feed, which book is recommended and which audience receives a particular promotion.
Even when an author creates every word and image manually, algorithms still influence whether that work is seen.
Are recommendation algorithms acceptable while generative tools are not?
Is automated advertising acceptable?
What about AI-assisted moderation, translation, captions, accessibility tools or spam detection?
These are not traps. They are genuine questions, and reasonable people may reach different conclusions.
When suspicion becomes gatekeeping
What concerns me most is when this moves beyond personal boundaries and becomes ideological gatekeeping.
Authors can be publicly “outed,” questioned or even excluded from signings because someone suspects AI may have been used in their writing, cover art, banners or marketing.
If boundaries are going to be imposed, they should be clear, evidence-based and applied fairly—not enforced through rumour, pressure or trial by social media.
Where are we prepared to draw the line?
Perhaps one author will reject AI-generated prose but continue using automated proofreading.
Another may accept assistance with marketing copy but never with a manuscript.
Someone else may use AI to generate an early concept that is later recreated by a human designer.
Another author may reject generative AI entirely while accepting that fraud detection, accessibility tools, recommendation systems and background automation are unavoidable parts of modern business.
Each of those positions can be discussed.
What becomes harmful is the assumption that every use is identical—or that an individual’s entire character can be judged from a single tool, image, phrase or accusation.
The author community often describes itself as gentle, supportive and kind.
At its best, it genuinely can be all of those things.
Yet some conversations around AI have become harsh, personal and accusatory. Authors are sometimes labelled dishonest, lazy or morally deficient before anyone establishes what tool was used, how it was used, or whether AI was involved at all.
That does not protect creators.
It creates suspicion and fear, while leaving the larger commercial systems—and the companies making many of the consequential decisions—largely unexamined.
A more useful conversation
A calm discussion about AI does not require us to approve of every use.
We can oppose the unauthorised use of copyrighted work.
We can ask for transparency about training data.
We can support human artists, editors, narrators, designers and other creative professionals.
We can require authors to remain responsible for the accuracy, legality and quality of what they publish.
We can make personal decisions about the tools we will and will not use.
Event organisers can also establish boundaries, provided those boundaries are clear, communicated in advance, evidence-based and fairly applied.
But we should be honest about the world in which those decisions are being made.
AI is already incorporated into websites, design programs, newsletters, payment systems, advertising platforms, editing tools, retailers and distribution services used by authors every day.
Sometimes we knowingly use an AI feature.
Sometimes we use the platform but avoid that feature.
Sometimes the technology operates in the background and we may not even know it is there.
There may be no single line that works for everybody.
So perhaps the question is not simply:
Do you use AI?
Perhaps the more useful questions are:
What kind of AI is being used?
What is it being used to do?
Was the source material obtained ethically?
Is a human still responsible for the result?
Has anyone been misled?
Has anyone’s intellectual property been exploited?
What evidence do we have before accusing someone?
And finally:
Where are each of us prepared to draw the line—and can we discuss those boundaries calmly, honestly and respectfully?
I do not expect everyone to agree.
I do hope we can talk about it without shame, suspicion or personal attacks.
A genuinely kind creative community should be able to hold strong ethical positions while still allowing room for evidence, complexity and respectful disagreement.
Be First to Comment