Author/Editor’s Note & TL;DR: This article has been updated to reflect the 2026 stance of the Authors Guild as well as to include some personal anecdotes about how social understanding has changed in the past year or so. I’ll likely to continue to update it as norms around AI use continue to evolve and as my job as a content marketer shifts in its presence. The TL;DR of my AI stance is that AI is not and never will be a creative substitute for authors and I do not approve of generative AI in creative spaces. In this article, though, I do reference assistive AI as a business tool. (So sorry for the clickbait headline lol.) As a marketer, I resist the common binary framing of “AI is evil” versus “AI is the future” and instead see the value of distinguishing between AI tools like Grammarly and AutoCrit versus things like ChatGPT-generated slop. While generative AI for writing, art, character development, and storytelling is ethically problematic, creatively inferior, and literally built on copyright-infringed training data, assistive AI applications can still provide value in non-creative small business tasks like marketing strategy, audience analysis, and website accessibility if used sparingly. Those who use it thoughtfully, transparently, ethically, and sparingly (for realsies, extra emphasis there) can save time while preserving their artistic integrity. The core message here? Reject AI-generated creativity; learn about AI-assisted productivity with a little side-eye to at least understand its opportunities and shortcomings.
There’s one important distinction to discuss before we get started here: AI isn’t a single technology or even a single category of tools. There are generative systems that produce new content, predictive systems that analyze existing data to anticipate future behavior, and automated systems that use machine learning to optimize repetitive tasks. These technologies can overlap, but they aren’t interchangeable, and they don’t all present the same ethical considerations. My focus here is primarily on the technologies that help businesses operate more efficiently without asking them to surrender their creative output.
Alright, with that out of the way, I’m going to get controversial. Generative AI sucks (especially in terms of what we can create by hand) and we can all agree on that. Using it for character art or language generation is ethically problematic, and the results? Glaringly unpleasant. AI-generated language and images do not belong in creative spaces and unabashedly steal from artists, which is something we very much need to push for regulation on. But I’ve seen many authors wholly condemn AI in general, and this is where the controversial part is going to come in… I believe we’re going to get left behind if we don’t at least recognize certain forms of AI as a tool. Why, you ask? Because the corporate world is already using it. Which means trad pub as an outlet is also using AI for authors’ marketing efforts, blurbs, and more. As a content marketer and corp comms professional, I’ve had to get certified in and comfortable with assistive AI use through a few different gigs, and I’m still employed as a working writer even though it exists. It simply cannot compete with the human element that goes into content creation – and, candidly, I’m skeptical that it will ever be able to replace that. Even if you were to nail down a formula for perfecting the exact elements of content creation, human thought cannot be replicated… And we all notice when it’s absent from a piece. With that being said, this new tech isn’t coming after our jobs, so let’s dive into how we can make the most of AI as authors.
Spoiler alert: I’m staunchly anti-generative AI in creative spaces. Assistive AI, though, has its place. That’s what we’ll be talking about here. When I say “assistive,” I mean administrative and analytical work — not writing books, creating characters, generating prose, making art, or replacing creative professionals.

AI as a Longform Tool
Once upon a time, an article fell on my desk. It was “written” by an agency partner at my day job, and it just wasn’t resonating with anyone on the team that requested it. Nobody could pinpoint what was wrong with it, but it wasn’t working. So they tapped me in to fix this monstrosity. After sitting with it for a bit, I definitely had a headache – it was confusing, heavy on the exposition, and it just didn’t flow.
And then I spotted the word “tapestry.”
Don’t get me wrong, I like the word. I’ve used it myself, but it’s one I’ve scrubbed from my vocabulary since I discovered that it’s a word that AI-versed readers recognize. AI LOVES that word, or at least one iteration of it did in the early days of wide adoption. Suddenly realizing this was an AI-generated article, I started analyzing what was missing. It had everything we had asked for in the story – all the product descriptions, all the value propositions we wanted included – but the content hierarchy did not follow the natural flow of conversation or human thought.
With a bit of restructuring and a heavy-handed rewrite, the article suddenly satisfied my stakeholders, including the c-suite.
What’s the moral of the story, and why am I choosing to lead with this before we go into how to use AI? It’s because I have seen firsthand how generative AI can follow a prompt and satisfy the request without satisfying the reader. In fact, that’s essentially what it’s trained to do. Satisfy the prompt. Even with high-tech and ever-adapting software, that human piece is absent, and readers notice.
So if you ever have the temptation to generate a scene… take it from me: I wouldn’t have a day job if AI was good enough to replace writers. I wouldn’t be trained in using it if it was likely to replace me. Instead, I’ve been taught to use it as a tool. The job wasn’t “replace the writer.” The job was “remove repetitive business tasks so the writer can spend more time writing.”
And that’s an important distinction in the professional world. A marketing department might use machine learning to identify which audience segments are most likely to convert, automation to distribute content, Grammarly to edit articles, and analytical tools to identify opportunities for improvement. None of those applications requires a machine to write the actual content. In fact, the more I work with these technologies, the more I see their value in what happens around the creative process rather than within it.
Let’s dive into that.
Ethical Use of Generative and Assistive AI for Authors
The first thing I learned to use AI for was generating SEO-rich alt text for images. What is alt text? It’s a piece of digital accessibility that’s essential for helping your visually impaired friends and search engines understand the images on a webpage. It’s absolutely essential and also rather annoying to write – and I say that as a visually impaired individual. (I lost my vision from iritis for several months a couple years back, and scarring has left me with some blind spots in my left eye. True story, but one for another time.) As writing alt text is such a breeze, I’ve not really adopted that practice long-term, but I still might use it to check for missing alt text, identify potential contrast issues, or detect improperly labeled form fields. What this all taught me early on is that if AI is going to be used at all, it should be used for non-creative applications. To me, that’s the ethical line. The moment AI starts replacing artistic decisions instead of administrative ones, we’ve crossed it. Since that first training, whenever I lean on AI as a marketer, it’s in a way that’s meant to save me time or take annoying tasks off my plate. Not to take creative endeavors from me.
And truthfully? That’s exactly why workplaces are pushing for its adoption – it’s a timesaver if used correctly, not a replacement for human creativity.

Assistive AI can be a strong tool for authors, and the Authors Guild has best practices outlined for us to reference.* They point us toward several ethical considerations – with the caveat that we recognize that many of these platforms operate using a large language model (LLM) “developed on mass copyright infringement.” (Bolded that because it’s a VERY important piece of the conversation. If you’re going to lean on any AI platform, do your research and understand how it was trained/continues learning.) With that in mind, you should not generate any language or art you’ll actually use. Personally, I would recommend not generating either ever, even out of curiosity… The results aren’t ever great, like that article I referenced earlier proved. Instead, you might consider using assistive AI as a tool for analysis and editing (AutoCrit and Inkshift are examples of AI editing software that performs tasks like crawling documents for word repetition, adverbs, and weak writing signifiers), according to the guidelines offered by the Guild.
*Edit in May of 2026: The Authors Guild has updated their AI guidelines to add context to the conversation. Once again, I find that I very much agree with their guidance. As outlined in this article where I attempt to define the difference between assistive and generative AI, the Guild states, “it is not useful to treat all AI use as ethically or professionally equivalent.” They also reinforce another major point I made earlier: AI cannot replicate human thought, and as such, it will never nail your unique voice. All it can do is produce a generic mashup of words it learned from mass copyright infringement. Again, it’s just never particularly good. I do encourage you to go read the Guild’s update, as it’s an excellent addition to the conversation. Ultimately, I’ll defer to the Guild again and quote, “Be aware of whether you’re displacing a human being when you use AI.”
This is where I think authors need to start looking beyond the generative AI debate. You can be entirely opposed to using AI to create your work while still recognizing that the businesses distributing, advertising, and selling that work are increasingly using machine learning and automated decision-making. Understanding that infrastructure doesn’t mean endorsing every application of it.
But let’s talk about applications. As I mentioned, I am a marketer. I’ve heard many of my peers suggest that if you’re going to use generative platforms for anything business-based at all, use them to explore marketing efforts. For example, you might give the platform you choose a high-level overview of what your story is about (and use placeholders for character names if you wish to preserve your anonymity and control) and ask it to generate ideas for a monthly social media calendar, or recommend audience targeting measures for whatever platform you’re running ads on. An important thing I’ve been taught is that due to the water toll of using AI, it’s much better to send fewer prompts, so ask it for a more comprehensive output (like a social media calendar, organized in a table with creative and image direction) rather than a smaller output (like applicable hashtags for a single social media post). With it doing high-level marketing brainstorming for you, you’re still operating as a creative agency… You’re just not operating without a brief anymore.
The important distinction here is that you’re using the platform to explore possibilities, not to dictate your creative direction. Think of it as an opportunity to get a preliminary marketing brief on paper. You’re still the creative agency. You’re simply giving yourself something to work from.
On the social media note, I do want to mention that it is just an example of how this type of application may be relevant, and it’s not necessarily something that is going to be particularly useful for everyone. I recently tested this particular concept as a potential method to fill gaps in my own calendars when I’m in a creative rut, and the results were rather interesting. [For the sake of full transparency, I used business CoPilot, which both my day job and a few clients lean on for marketing work. This uses enterprise security and does not use data to train other AI models, and prompts are discarded when sessions end.]
These are some of the ideas it provided, based on a prompt that informed it about a dystopian romance manuscript with pirates, biopunk elements, slow-burn pacing, and high stakes in a lore-rich world:
- “Meet the Crew” Carousel – Include aesthetic moodboards for each key character.
- 👉 Add a “Which character are you?” quiz link in your stories for bonus interaction.
- Slow Burn Snapshot – Use a dramatic or dreamy image of a single bed or a half-empty glass paired with some teaser snippets from your manuscript.
- Things [Your Character] Would Never Say Out Loud – Gritty or emotionally charged black-and-white portrait with a conceptual quote.
- Lore Teaser: Grungy iconography or geometric symbols couples with worldbuilding details.
Around the time those were generated and this article was written, I thought these were interesting suggestions. I saved them, figuring they’d be useful if I ever found myself having a particularly slow day and needed inspiration. One prompt had generated an entire document of potential content, ostensibly enough to fill gaps in my calendar for a year or longer.
But here’s where things get interesting.
One year later as I’m editing this article in 2026, I haven’t used a single one of these ideas. In fact, since this experiment, I haven’t leaned on AI for my author brand’s social media work at all. My post ideas continue to come primarily from conversations with readers and other authors, and the suggestions Copilot generated simply never made it into my creative process.
I don’t think that makes the experiment a failure, necessarily. It demonstrated that generative AI can produce relevant suggestions when given a sufficiently detailed brief. But relevance isn’t the same as resonance.
Interestingly, I did find myself ideating beyond these suggestions, using them as fuel for my own creative juices. A lore-teaser moodboard carousel, for example, would be a genuinely fun project. But the inspiration to execute it never materialized from the generated document. That creative impulse still had to come from me.
You’ll also notice that none of these generated suggestions include captions or images. There’s a reason for that… Like we discussed early on, AI text satisfies the prompt, not the reader. AI-generated captions, on-image text, and art remove the human element and only serve to isolate your audience. If you’re going to lean on it for anything close to creative work, put it in an assistive role – have it essentially fill out a marketing brief, then leave the actual creative heavylifting to yourself. And don’t treat this as a firm marketing strategy. Your audience, your creative instincts, and your understanding of your brand should remain the deciding factors.
Which actually leads me into another point I’d like to mention – if you’re going to use AI, use it sparingly. Seriously consider only using it for asks that will serve you for ages and keep those creative juices flowing. And if you are considering using it for more creative tasks, research what Fairly Trained certifications are and why they’re important. [Full transparency again: I’ve never used it for more creative tasks and don’t love the idea of AI touching creative work, so I can’t speak to these particular platforms. Just passing on resources from my author groups, so seriously, do ya research.]
And, again, on the note of using it sparingly… One ChatGPT query is comparable in energy usage to 10 Google search queries. That’s also comparable to the amount of energy required to keep a light bulb on for 20 minutes. Whether you’re ChatGPTing or Google searching your day away, you might consider turning to old-fashioned bookish research if you’re looking to lessen your environmental footprint. And I’ll also add that if you’re using AI often enough that it’s becoming part of your creative routine instead of an occasional business assistant, I’d argue it’s worth reassessing whether it’s actually helping or simply replacing work you would have grown doing yourself.
(Edit in May of 2026: I still stand by my assertion that using electronics period result in an environmental impact, and I think that’s become an even bigger piece of conversation in the time that’s passed since… Not only should we be holding our respective governments responsible for regulating AI and the impact it has on the energy grid and other resources, but we should also have them regulate the impact of Google searches, Teams meetings, frickin’ Bitcoin transactions, and everything in between that relies on data centers. Let’s demand that all these tech giants invest in making ALL their technologies less impactful across the board. That said, as more research has surfaced about AI query energy use, it’s evident that we truly don’t have exact data. There’s a lack of transparency in how calculations like the one above were made, and there’s no real standardized methodology for calculating impact. Ultimately, in the past year, I’ve only grown more convinced that we need to regulate and force AI providers to invest in year-over-year energy efficiency. As much as I wouldn’t mind seeing it go away, I don’t believe that will happen anytime soon and feel very strongly that we need to make it less environmentally impactful in the meantime. We the people need to hold our government accountable in enforcing these regulations, because big businesses won’t invest big bucks in something if they don’t have to. Like, even if we personally hate it and never use it and continually educate about its impact, AI isn’t going fully away. So let’s work together and fix that pesky little impact piece.)
AI Usecases Authors Might Consider
Of course, generating usable outputs requires a strong prompt and, potentially, a bit of in-platform training. If you’re new to writing prompts, I’d like to give you the easiest direction imaginable – start simple. Remember that the goal of using AI as an author or marketer is to save time. And if you’re going to use it for any sort of analysis, remember that these platforms are trained both to satisfy the prompt and you. We’ll get to that in a sec.
If you’re looking to generate a prompt, especially in terms of marketing activities, consider some of these tips:
- Start with a strong prompt foundation. Through a previous day job, the big ol’ corporation pulled some strings and convinced their Microsoft rep to give the marketing department training in using CoPilot “the right way.” Now, if I use AI at work, I usually start with something like, “I’m a content marketer looking to expand my organic marketing strategy for a home automation system that puts thermostat, lighting, and blind-lowering control in a homeowner’s hand. [Here’s what I am asking for.]” In much simpler terms, you could say “I’m a [title]. I want you to [perform a task] in [output format].” “I’m a web builder and author. I want you to create some code for a foldable accordion with six sections and populate it with lorem ipsum in a format that I can copy and paste into my website.” Boom, one query that will serve you for ages. If you’re gonna use it for assistive generation like this, use it sparingly. As an important reminder, AI cannot reliably identify what makes your particular creative voice, brand, or relationship with readers distinctive. That means that when you use it beyond tasks like this to do things like creating a brief, it should support your creative process rather than substitute for it.
- Create some content pillars. As an author, you’re a brand. You know what brand builders love? Content pillars. If you have a website or web-based presence like a blog, you can ask AI platforms to crawl your website and analyze your content pillars. These can inform content strategy for blog posts, social media calendars, and even adjacent genres you might consider writing in. For the sake of this article, I had ChatGPT crawl this very site for its content pillars. (Again, I’m a content marketer, and I’ve been writing professionally since 2007. I’m certain my historic content has already been used to train AI, and this website’s very existence likely invites crawling. So… I’m approaching this with the assumption that it’s already been crawled in mind.) It highlighted Speculative Fiction with Social Commentary, Fantasy with Mythological Elements, Writing Craft and Author Resources, and Indie Author Community Engagement as my brand’s content pillars. I see a little bit of those elements in my social media presence, but definitely not enough.
- Explore tropes and target audience information. I’ll be the first to say that I’m a content marketer, not a media buyer. I write ads for the audiences our media buyers tell me to. As such, I’ve used things like Hubspot’s persona AI to analyze rough in-the-moment summaries of my books to narrow in on my target readership persona for marketing purposes. (And I do mean rough, excluding major plot points and just giving it a loose outline of what I consider the themes to be and maybe a one sentence elevator pitch.) According to AI, adult SFF readers hungry for deep, character-driven plots and political dramas with teeth, not just aesthetic are my ideal readership. Readers of “more angst, less smut” who like literary-adjacent literature will resonate with my work. Now, that’s one hell of a broad audience, but creating targetable personas and exploring comp titles from there can help tremendously with marketing efforts.
- Generating newsletter topic ideas. I keep talking about marketing, but let’s discuss a real-life digital marketing use case. Your readers may be sick of those “book update” emails, but working off your recent blog posts, social media posts, or social appearance updates can inspire ideas that will keep your customer base hooked.
Notice that every example in this article is about planning, marketing, organization, or analysis… not creating art. That’s very much intentional. And while we’re here, let’s address a few things you should never, ever let AI touch:
- manuscripts
- prose
- dialogue
- scenes
- characters
- worldbuilding
- cover art
- interior illustrations
- audiobook narration
- book translations
- anything representing your creative voice
- commissioned artwork or other work that should be performed by a human creative professional
With all these items in mind, always remember to be careful with prompt bias. You think your story is good, but seriously, if you’re using AI for any form of analysis… Request brutal honesty if it’s prompt-based. If it’s an automated tool like AutoCrit or Inkshift, take everything it says with a full pound of salt. By that, I mean don’t let its pre-programmed praise get to your head. I’ve had many freelance writers reject critiques from me as an editor because they asked AI to explain why their writing is good when their corporate client told them it isn’t. Bias exists, both in prompts and generated content, so stay diligent in watching for that.
Of course, not every AI application involves prompts. One area where authors may already be benefiting from AI without realizing it is paid advertising. Platforms like Meta and Google increasingly use machine learning to optimize ad delivery, audience targeting, and bidding. Meta’s Advantage+ tools, for example, automate aspects of audience targeting and campaign optimization, while Google’s Smart Bidding evaluates contextual signals like search queries, device, and time of day to determine how much to bid for individual ad placements. You might also be leaning on AI like:
- Website analytics and audience behavior – If you have an author website, Google Analytics 4 offers predictive metrics that can help you better understand visitor behavior. Depending on the amount and quality of data your website collects, its machine-learning models can estimate things like purchase probability, churn probability, and predicted revenue. In practical terms, this can help you identify patterns in visitor behavior and understand which audiences are more likely to take a particular action. Now, a quick reality check I just have to include: these features aren’t particularly useful for a brand-new author website with a handful of visitors. But if you’ve been building your platform for a while and have a meaningful amount of traffic, it’s worth exploring.
- Paid advertising optimization – I know I just mentioned that paid advertising uses machine learning, but I want to explore this a bit deeper and encourage you to consider what this actually means for an indie author. Let’s say you’re advertising a new fantasy release. You might have a general idea of your ideal readership, a budget, and a few creative assets you’ve commissioned from an artist. You can provide the platform with your campaign objectives and audience parameters, then allow its optimization systems to evaluate performance and adjust delivery. You’re essentially allowing the platform to use its data to make more informed decisions about where and when your advertisement appears. These systems, though, are only as useful as the objectives you give them. If you’re optimizing for clicks, the platform will look for people likely to click. That doesn’t necessarily mean they’re likely to purchase your book. If your ultimate goal is sales, make sure you’re tracking meaningful conversions and not simply celebrating an inexpensive cost-per-click. Automation can optimize your campaign, but it can’t compensate for an improperly configured marketing strategy.
- Newsletter delivery – If you’re using an email marketing platform like Mailchimp, you may have access to features that analyze subscriber engagement and help determine when to send your campaigns. Mailchimp’s Send Time Optimization, for example, uses data science to identify when individual recipients are most likely to open an email within your selected 24-hour delivery window. Essentially, you’re still writing your newsletter, choosing your announcements, and deciding what your readers need to hear. The technology is simply helping you get that human-created content in front of them at a more appropriate time.
- Audience segmentation – Beyond delivery times, behavioral segmentation can help you understand how different groups of subscribers interact with your business. For example, you might have readers who consistently click through to preorder announcements, subscribers who only engage with cover reveals, and people who primarily visit your website for writing advice. Your email platform can likely help you organize these audiences based on their actual behavior, allowing you to make more informed decisions about what information to send and to whom. If it doesn’t, it likely has served you enough data to turn around and ask the platform of your choice to analyze it.
- Search and website optimization – Google Search Console, for example, can help you identify which search queries are bringing visitors to your website, which pages are receiving impressions, and where your content might have opportunities to perform better. While not every feature in platforms like Google Search Console is AI-powered, they demonstrate the broader point: you can use technology to understand how readers discover your work without asking it to create that work for you.
In practical terms, this all means you don’t necessarily need to know every characteristic of your ideal readership to get started with paid advertising. These platforms can do what they do (analyze performance data and identify patterns that help them deliver your ads to people more likely to take action), but you’ll still be responsible for your campaign objectives, budget, messaging, and creative direction. But with their built-in AI tools, the platforms you’re advertising on can help with the analytical heavy lifting.
And importantly, none of this requires you to generate a single piece of creative content with AI. Like I said in the beginning… As a marketer, I simply don’t think every form of AI is evil. It’s impossible for me to escape it my job, and after it being around for so long, I’m not scared of it. The goal of this article is to help you feel less threatened by it, too.
And since we touched on generative applications in this section, I have to mention… There’s an important difference between a platform identifying patterns in your actual audience data and a generative model producing a hypothetical customer persona based on a book description. The latter can be an interesting thought exercise, but it isn’t market research. Don’t mistake a convincingly written description of your ideal reader for evidence that those readers actually exist, behave that way, or are interested in purchasing your work.

Author Disclosure of AI Use
If you’re going to use AI, disclose that fact when it matters. Not because someone’s going to slap your wrist for using AI to brainstorm which tropes apply to your story, but because it can actually impact contracts and copyrights.
Readers want to know that you are the voice behind the story. If you use AI to generate scenes or dialogue (which, again, I really don’t recommend or find to be honest), that’s a different conversation… According to the Authors Guild, AI-generated characters, plotlines, and text can violate contracts. If there is AI content in your story, your agent or publisher will need to know upfront so they can properly register and protect your manuscript.
But if you’re very sparingly using AI like I do – to brainstorm content marketing briefs for future use or come up with audience targeting measures every other quarter or so – you still preserve the integrity of your work and ideas. My philosophy is simple: if removing AI from the process would fundamentally change the creative work itself, I don’t think AI belonged there in the first place. Let’s be honest… Most of us write because we love it, so it’s important to hold tight to the preservation of our efforts. And it’s also equally important to continue supporting our fellow artists, whether they’re character artists or fellow authors. It’s absolutely essential that we act as a community to support and celebrate what human artists add to the world.
Above all else, the Authors Guild encourages transparency, and I agree. But I also think we can be nuanced. A footer that says “AI was used to draft this email” is not necessary if the only thing the AI touched was subject line themes you only used as inspiration. Use common sense. Be honest. Don’t let tech take the wheel.
You’re still the one driving.
AI Won’t Replace Us Artists
At the end of the day, I don’t believe AI can replace what authors uniquely contribute – but it does have value when used to assist us from a business and marketing perspective. Is there an economic incentive to using it for things like cover generation? Sure, it would certainly be cheaper monetarily, but the actual cost could be your entire publishing career given how staunchly anti-AI the community is. But limited, non-creative, almost administrative use of agentic and assistive tools truly can have value. In summary, reject AI-generated creativity. Learn enough about assistive AI to understand where it genuinely helps and where it absolutely doesn’t.
Whether you’re writing newsletters, crafting paid social media campaigns, adjusting code for your website, or just trying to decide the best day to launch your book, AI can be the tool that helps you get there more efficiently. If the right prompt saves you ten Google searches, more power to you… But invest some time in training yourself so you fully understand what this tool actually is and why you shouldn’t just use it every day for everything. It’s not magic. It’s not evil. It’s a glorified autocomplete that sometimes spits out a genius idea and sometimes suggests the word “tapestry” 47 times in a single draft. But we can take a deep breath… Even the most brilliant ideas it provides can’t replace a single drop of your creativity. And believe me, readers notice. Your authenticity is more valuable than what the internet has appropriately dubbed “AI slop.” Every. Single. Time.
And perhaps that’s the biggest takeaway I hope authors get from this article: you don’t have to embrace generative AI to understand the technology that’s changing the business side of publishing. You can refuse to use AI-generated artwork, reject machine-written prose, and continue commissioning human creatives while still recognizing that the platforms helping you reach readers and market to them are evolving. Understanding those systems doesn’t make you any less of an artist. It makes you a more informed business owner.
So let’s stop pretending AI is the end of creativity as we know it. You are invaluable, and your stories deserve to be told. Avoid generative AI. Use assistive AI wisely. Use it sparingly. Use it with a little side-eye. Protect your craft, but don’t be afraid of where the future is taking us.
Keep writing like only you can.