Artificial intelligence (AI) dominates today’s headlines, often swinging between hyperbolic promises and existential fear. Depending on what you read, AI is either poised to solve humanity’s greatest problems or destroy creative industries altogether.
The reality is more nuanced.
For authors, AI is neither wholly good nor wholly bad. It is a powerful toolset that introduces both creative opportunities and ethical risks. Understanding where those opportunities and risks lie is becoming increasingly important for writers navigating today’s publishing landscape.
While this article is not legal advice and specific legal questions should be directed to an attorney, I approach this topic from the perspective of someone who works in and deeply cares about AI ethics. My opinions are my own and my intention is to help authors navigate this terrain responsibly.
What is AI, anyway?
To understand the current debate around AI, it helps to briefly understand how we arrived here.
Researchers began exploring artificial intelligence in the 1950s, attempting to build systems that could mimic aspects of human learning and reasoning. In those early days, AI was considered by many computer scientists to be a fringe field of study.
Over time, breakthroughs in machine learning, reinforcement learning, and deep learning laid the groundwork for neural networks capable of recognizing patterns and generating output. Earlier forms of AI already power many familiar tools, from predictive text and recommendation engines to fraud detection and search algorithms.
The breakthrough that captured the public’s attention came with the commercialization of generative AI through the launch of ChatGPT in 2022. Large language models (LLMs) generate responses by predicting the next likely token, or fragment of language, based on patterns learned from enormous amounts of data. The results can sound remarkably knowledgeable, persuasive, or creative. But beneath the fluency, these systems are still probabilistic prediction engines shaped by their training data, software architecture, and human-designed guardrails.
Since then, AI capabilities have expanded rapidly. Multimodal systems can now generate text, images, video, and audio. Some models can imitate voices or likenesses of real people, which has fueled concerns about misinformation and “deepfakes” when used deceptively. Newer reasoning-oriented models can display step-by-step processes intended to help users understand how outputs are generated. Meanwhile, agentic AI systems are beginning to autonomously complete tasks on users’ behalf. The next frontier may be physical AI embedded into robotics, vehicles, and wearable technologies.
Sensational or sensationalized?
The public conversation around AI is often shaped by headlines that amplify excitement or fear. On March 16, 2026, the New York Post published a story claiming that a “Tech pro saves his dying dog by using ChatGPT to code a custom cancer vaccine.” Stories like this quickly fueled online searches suggesting that “ChatGPT cures dog cancer.” The reality was more measured. The dog owner collaborated with a medical professional and used AI as one component of a broader problem-solving process. The dog’s tumor reportedly shrank significantly, which is certainly noteworthy, but the story does not mean AI independently cured cancer.
Likewise, headlines surrounding advanced AI systems can lean toward sensationalism. For instance, reports about Anthropic’s guarded “Mythos” model initially indicated that unauthorized users accessed the locked-down model. Yet, later media stories reported that the intruders appeared more interested in experimenting with the system than unleashing catastrophic cyber attacks. This tension between hype and reality matters because it shapes how writers and readers think about AI’s role in creative work.
AI for authors
Given this context, it should not be surprising that AI’s use in writing is neither entirely beneficial nor entirely harmful. Rather than thinking about AI as simply “allowed” or “forbidden,” it may be more useful to think about it as a spectrum of creative delegation. If you choose to leverage it, AI can help authors brainstorm ideas, organize research, refine marketing copy, generate social media concepts, or assist with editing. Used thoughtfully, these capabilities can save time and unlock new creative possibilities. Alternately, it is every author’s prerogative to forgo using AI. For those using AI, the technology introduces meaningful risks that authors should understand.
Risk: hallucinations
Generative AI systems can produce inaccurate or fabricated information, often called “hallucinations” or confabulations. Because these systems are designed to generate plausible language rather than verify truth, they can present false information with remarkable confidence.
For nonfiction or historical writing, this can create obvious problems if authors rely on AI-generated research without independent verification. On the other hand, for brainstorming story ideas or exploring creative possibilities, the same pattern-generation capability can sometimes be useful.
The key distinction is recognizing where factual precision matters and where imaginative exploration is appropriate.
Risk: bias
Bias is another important concern. AI systems reflect both the strengths and flaws of the data used to train them, as well as the guardrails built around their outputs. This matters when authors use AI for research, characterization, visual generation, or audience analysis. Harmful stereotypes, cultural blind spots, or unsafe outputs can emerge if models are poorly trained or insufficiently moderated. As with any research source, authors should apply critical thinking rather than treating AI output as inherently objective or authoritative.
Human in control
AI systems can generate polished and persuasive output that sounds authoritative even when inaccurate. Because of this, authors should view AI as an assistant rather than an authority.
Human judgment remains essential, particularly for factual research, sensitive subjects, ethical decisions, and final creative direction. The more responsibility we delegate to machines, the more important human oversight becomes.
Risk: copyright and creative ownership
Questions surrounding copyright, intellectual property, and training data remain active areas of legal and ethical debate. Ongoing lawsuits against AI companies have raised important questions around fair use, creator consent, and how training materials were obtained. Legal standards continue to evolve rapidly, which is one reason authors should stay informed as the landscape changes. These concerns extend beyond text generation to images, video, and audio. For example, many authors feel uneasy about using AI-generated cover art because some image-generation systems may have been trained on copyrighted creative work without artists’ permissions. In addition, in the United States, fully AI-generated text or visuals may face challenges receiving copyright protection unless sufficient human authorship is involved.
Audio raises similar concerns. Using AI-generated replicas of real people’s voices without explicit permission can create ethical and legal risks, particularly when audiences may believe the voice is authentic.
Risk: sustainability impact
The energy, water, and material impacts of digital, data, and AI is a larger topic than this short article can cover. In general, we should be aware that a larger proportion of AI’s impact comes from training the models than any given prompt. Thus, the ways we can positively affect this area of risk likely comes down to holding tech companies who are creating next-generation models accountable to their ESG (environmental, social, governance) commitments and avoiding frivolous activities in our own use. AI models are powerful pattern recognition machines, so use them for complex projects but don’t engage them to do simple things you can find elsewhere or to create non-productive AI slop.
Risk: trust
Trust is critical in every relationship. Because of the issues above and more, some people identify as anti-AI. Now, with more information, you can make an informed decision about where your beliefs lie, consciously choose where you will and won’t use AI, and be transparent in disclosing its use.
Opportunities
While it can seem easy to point to the risks, it’s also important to acknowledge the opportunities that AI can provide in accessibility, creative support, and productivity. For example, AI has the potential to democratize books through language translation, disability accessibility, and marketing support for authors with limited resources.
When does AI stop assisting and start authoring?
Because of the way that AI is trained on human-created works, many authors feel comfortable using AI as an editorial or brainstorming assistant while remaining deeply uncomfortable with books that are largely generated by AI. Likewise, readers generally expect that books are based on human imagination, storytelling, emotional experience, and writing. The ethical concern is not simply whether AI was used, but whether the human author remained meaningfully engaged in the creative process. Many in the writing community view fully AI-generated books differently from AI-assisted writing because the acts of choosing language, shaping scenes, and wrestling with meaning have traditionally been central to authorship itself. This raises philosophical as well as practical questions:
- If AI generates most of the prose, who is the author?
- Is prompting equivalent to writing?
- What level of human revision meaningfully transforms AI output into original creative work?
- What expectations do readers reasonably have when they purchase a book marketed under a human author’s name?
Practical guidelines for authors
Transparency and intentionality can help authors navigate these choices responsibly.
Some practical considerations include:
- Disclose substantial AI use. Readers and publishing professionals increasingly value transparency. If AI materially contributes to your finished work beyond light editing or brainstorming, disclosure helps build trust.
- Understand the tools you use. Not all AI companies approach training data the same way. Some companies, including Adobe Firefly, Getty AI, and Shutterstock, publicly state that they train on licensed or permissioned creative content. Personally, I gravitate toward tools that have committed to ethical sourcing for their training.
- Keep humans meaningfully involved. AI can accelerate parts of the creative process, but authors remain responsible for accuracy, ethics, originality, and final judgment.
- Avoid deceptive uses. Be especially cautious with generated images, cloned voices, or realistic video that could mislead audiences or imitate real individuals without consent.
The creative continuum
One useful way to think about AI is along a continuum measuring how much humans drive the creative process versus how much is delegated to machines.
Not every point along this continuum carries the same ethical weight.
For example:
- Many readers may feel comfortable with authors using AI for brainstorming or editorial assistance.
- AI-generated social media posts may warrant disclosure depending on how extensively they are used and whether there could be risk of deception (for instance, fake content that looks realistic).
- AI-generated cover art or audiobook narration may raise more significant concerns around artistic ownership, transparency, or authenticity.
- Fully AI-written books fundamentally change reader expectations about authorship itself. I myself err on the side of human-driven, using AI for ideation or editing support rather than writing.
The more AI materially contributes to finished creative work, the more important transparency and human oversight become.
Social media provides a useful example of the judgment calls authors increasingly face. If AI-generated content could mislead readers, spread misinformation, or substantially replace authentic communication, disclosure is often the wisest course. When in doubt, transparency usually strengthens trust rather than weakening it. The publishing industry itself is still actively navigating these questions. One publisher recently withdrew a book suspected of heavy AI authorship, suggesting that many readers and publishers place significant value on human-created storytelling. Companies like Amazon KDP require publishers to disclose whether AI-generated content was used during book creation.
You can certify your writing as human-generated by applying for a “Human Authored” mark through The Author’s Guild or companies like Verify My Writing. Note, though, that AI-detection tools aren’t fool-proof.
Summary
AI is not replacing the need for human creativity, judgment, or ethics. But it is reshaping how creative work can be developed, refined, and distributed. For authors, the important question is not simply whether to use AI, but how intentionally and transparently to use it. Some writers may choose to use AI sparingly for brainstorming or editing support. Others may embrace it more extensively for marketing, imagery, narration, or drafting. Wherever you fall on that continuum, thoughtful human oversight remains essential. Ultimately, readers connect not just with polished output, but with authenticity and human perspective. Those qualities remain uniquely human, and uniquely valuable.
________________________
Carol Van Den Hende is the award-winning author of the Goodbye, Orchid series, which draws from her Chinese American heritage and has won 40 literary and design awards. She’s also an MBA with decades of experience in marketing, strategy, and insights who is passionate about simplifying marketing concepts into actionable steps for publishing success. She’s keynoted and presented at conferences like Writer’s Digest, IBPA, International Women’s Writing Guild, Rutgers Writers’ Conference, Sisters in Crime, and Women Who Write. Carol’s mission is unlocking optimism as a writer, speaker, strategist, board member, and Climate Reality Leader. One secret to her good fortune? Her humorous hubby and twins, who prove that love really does conquer all.