AI as a Research Tool for Authors: A Real Example
Tuesday, August 11, 2026
Let me tell you about a client I'll call Sarah. She writes epic fantasy, the kind with intricate world-building, morally complicated characters, and battle scenes that make you wince in the best possible way. She came to me frustrated. She had spent the better part of an afternoon down a Google rabbit hole trying to figure out how a medieval healer would treat a stab wound, and she had come up with nothing useful. Contradictory forum posts, Wikipedia articles that cited other Wikipedia articles, and one very enthusiastic blog about LARP costuming that was absolutely no help whatsoever.
So I said, casually, "Have you tried asking ChatGPT?"
You would have thought I suggested she sacrifice her manuscript to a dark god.
"But That's AI. I Can't Use AI."
The reaction was immediate and, I will say, impressively theatrical. Sarah was convinced that touching anything AI-related would get her blacklisted from every writing community she belonged to, that her book would be labeled AI-generated, that readers would somehow know, and that her entire career would implode before her second cup of coffee cooled.
I let her finish. Then I asked her a question: "When you Google something for research, are you using an algorithm to find information?"
She paused. "That's different."
Is it, though? We spent the next hour talking about what AI actually is as a research tool, as opposed to what the loudest voices in online writing communities tend to conflate it with. Because here is the thing that gets lost in all the noise: there is a profound difference between asking an AI to write your book and asking an AI to tell you how medieval wound care worked.
One is outsourcing your creativity. The other is doing research. Efficiently.
Fine. Let Me Show You.
Rather than continue the debate in the abstract, I opened ChatGPT and typed in a prompt right there, live, while Sarah watched. The prompt was straightforward and specific, the kind of thing any fiction author doing their homework would ask. Here is exactly what I typed:
The Research Prompt
"If someone is stabbed in the middle ages, what would be the process to treat the wound. What natural compounds would be used to help the blood clot, to prevent infection and to speed healing. I'm writing a fantasy based around middle age or medieval technologies and I want to make it sound realistic. How long would healing take for a typical battlefield injury like a slash with a sword, and how about a simply stabbing, like in a street fight."
And then we watched what came back. What followed was exactly the kind of thorough, organized, historically grounded information that Sarah had been hunting for all afternoon. I have reproduced it in full below, reformatted slightly for readability, but with all the content intact. Read it as a writer. Ask yourself whether this is any different, in kind, from what a medical historian would tell you over coffee.
The ChatGPT Response
The following is the full, unedited response from ChatGPT, reformatted for readability.
ChatGPT Response
Medieval Wound Treatment: A Guide for Fantasy Writers
Great question for world-building! Here is a detailed breakdown of how a medieval healer or battlefield surgeon would treat a stab or slash wound, using period-appropriate methods and natural medicines.
Stage 1: Stop the Bleeding
- Direct pressure:Cloth, leather, or bare hands pressed firmly against the wound. Often held for a long time.
- Elevation:If a limb, raise it above the heart.
- Tourniquet:A strip of cloth or leather tied tightly above the wound on a limb. Medieval healers understood this worked, even if they did not fully understand why. Prolonged use could cost the limb, which they also knew.
- Cauterization:A heated blade, iron, or even boiling oil poured into the wound. Brutal, painful, and used mainly for amputations or wounds that would not stop bleeding. It was a last resort, not a first choice.
Stage 2: Cleaning the Wound
Infection was the real killer, and medieval healers knew it, even without germ theory. They used:
- Wine:The alcohol content made it a reasonable antiseptic. Common, widely available, and frequently used to wash wounds.
- Boiled water:Not always standard, but some healers, particularly those influenced by Islamic medicine (which was significantly more advanced), used clean boiled water.
- Vinegar:Acetic acid is a decent antimicrobial. Used to clean wounds and sometimes soaked into dressings.
- Salt water:Saline solution, essentially. Painful, but effective at drawing out fluid and cleaning.
Stage 3: Remove Foreign Material
Bits of cloth, dirt, metal fragments, and bone splinters had to come out. Healers used probes, forceps, and their fingers. Leaving debris behind was a near-guaranteed death sentence from infection. This step was often the most agonizing part of treatment.
Stage 4: Stitching or Binding
- Sword slash (surface wound):If the edges were clean, they might be stitched closed with thread (linen, silk, or even horsehair) or held together with strips of cloth or leather. Egg white was sometimes used as a kind of adhesive sealant.
- Deep stab wound:Often left partially open or drained deliberately to prevent pus from building up inside. Closing a deep wound too tightly could trap infection. Healers sometimes packed the wound with medicated cloth to encourage drainage.
Natural Medicines Used
Here is where your healer character gets their kit:
- Honey:Raw honey is genuinely antibacterial, hygroscopic (draws moisture out of bacteria), and was widely used. It also kept dressings from sticking. This is not folk superstition. It works.
- Garlic:Contains allicin, a natural antimicrobial. Applied directly to wounds or used in poultices. Smells terrible. Effective.
- Spider webs:Contain vitamin K, which helps clotting. Applied directly to a bleeding wound. Gross but historically documented and actually somewhat effective.
- Yarrow (Achillea millefolium):Named after Achilles, who supposedly used it on his soldiers. A powerful styptic (stops bleeding) and mild antiseptic. Applied as a poultice or packed directly into a wound. One of the most important battlefield herbs.
- Plantain (Plantago major):The common weed, not the banana. Anti-inflammatory, draws out infection, promotes healing. Chewed or crushed and applied as a poultice.
- Comfrey:Called "knitbone" for good reason. Contains allantoin, which accelerates cell regeneration. Applied to wounds to speed healing. Excellent for bruising and deeper tissue damage.
- Willow bark:Contains salicin, the precursor to aspirin. Used for pain and fever reduction. Given as a tea or decoction.
- Calendula (pot marigold):Anti-inflammatory, antifungal, promotes skin healing. Used in salves and poultices.
- Sage and Thyme:Both have antimicrobial properties. Used in washes and poultices.
- Lavender:Antiseptic and calming. Used in wound washes and to help the patient sleep.
- Moss (particularly sphagnum moss):Used as wound packing. Naturally absorbent and mildly antiseptic. Used as recently as World War One for exactly this purpose.
Dressings
Clean linen was the gold standard. Dressings were changed regularly and checked for signs of infection. Honey or herbal salves were often applied between changes. Egg white, as mentioned, was used as a sealant on surface wounds.
Pain Relief
- Willow bark teafor ongoing pain and fever.
- Alcohol(wine, ale) to dull the senses before and after treatment.
- Opium poppywas known and used, though access varied by region and wealth.
- Henbane and mandrakewere used as sedatives and anesthetics, though dosing was dangerously imprecise. A skilled healer knew how to use them. An unskilled one could kill the patient.
Signs of Trouble
Medieval healers watched for these, even without the vocabulary of modern medicine:
- Increasing redness, swelling, and heat around the wound (inflammation and early infection)
- Pus: thin yellow pus ("laudable pus") was considered a good sign in medieval medicine, a sign the body was fighting. Thick, foul-smelling green or black pus was a very bad sign.
- Red streaking spreading from the wound (septicemia, blood poisoning)
- Fever, chills, confusion (systemic infection)
- Gangrene: blackening of tissue, foul smell. Amputation was the only option, and survival was not guaranteed.
Healing Time: Realistic Timelines for Your Story
This is where fiction often gets it badly wrong, either healing too fast or too slow. Here are realistic ranges assuming no major complications:
- Minor slash (surface cut, clean edges, no muscle damage):Skin closure in 5 to 10 days. Full healing 2 to 4 weeks. Character can be mobile but should be favoring the wound.
- Moderate slash (deeper, muscle involvement, significant blood loss):Initial stabilization 1 to 3 days (bed rest essential). Wound closure beginning by day 7 to 10. Back on feet (carefully) in 3 to 4 weeks. Full strength return in 6 to 8 weeks. Risk of reopening wound with exertion is real and should affect your plot.
- Deep stab wound, street fight (clean blade, no organ damage, torso or limb):Most dangerous period is the first 72 hours. Infection risk is high. If the patient survives the first week, prognosis improves significantly. Bed rest for 2 to 3 weeks minimum. Light activity by week 4. Full recovery in 2 to 3 months, with residual stiffness and pain likely permanent.
- Abdominal stab (gut wound):Historically, a near-death sentence. Even a small puncture to the intestine releases bacteria into the abdominal cavity. Peritonitis sets in within 24 to 48 hours. Without modern surgery, survival rate was very low. If your character survives this in a fantasy setting, there should be a reason (exceptional healer, magical intervention, extraordinary luck). Recovery, if it happens, would take months.
- Chest stab (lung involvement):A collapsed lung (pneumothorax) was treatable if a healer knew to create a chest drain to release trapped air. Survival was possible. Recovery was slow, 2 to 4 months, with reduced lung capacity likely permanent.
A Battlefield Sword Wound: Day by Day
For a significant slash wound received in battle, here is what a realistic timeline looks like:
- Day 1:Bleeding controlled. Wound cleaned. Patient in shock, possibly unconscious. Fever may begin.
- Days 2 to 3:The critical window. Infection either takes hold or does not. High fever, pain, possible delirium. This is when characters die in realistic fiction.
- Days 4 to 5:If fever breaks, cautious optimism. Wound begins to close at edges. Patient exhausted, weak, in pain.
- Days 7 to 10:Stitches or bindings can be assessed. Wound should be knitting. Patient can sit up, take broth, have short conversations.
- Weeks 2 to 3:Patient can move around slowly. Wound is fragile. Scar tissue forming.
- Weeks 4 to 6:Light activity. The wound pulls and aches. Scar is visible and sensitive.
- Months 2 to 3:Functional recovery. Strength returning. Scar permanent.
For a Fantasy Setting: Your Healer's Kit
If your healer character is carrying a kit, here is what would be in it and what each item does:
- Linen bandages (clean, rolled)
- A small pot of raw honey
- Henbane or poppy preparation (carefully sealed, used sparingly)
- Calendula salve (in a small sealed pot)
- Dried yarrow (loose or in a pouch, for direct wound application)
- Dried comfrey (for poultices)
- Willow bark (for pain and fever tea)
- A vial of wine or vinegar (for cleaning)
- Needle and thread (linen or silk)
- A small knife and probing tool (for removing debris)
- Sphagnum moss (for packing and absorption)
- A pouch of salt
One final note worth remembering for your writing: the outcome of a wound in the medieval period had less to do with the size of the weapon and more to do with what structures it damaged, whether the wound was cleaned promptly, and whether the patient had access to a competent healer. A small knife wound to the abdomen could kill in three days. A dramatic slash across the chest might heal cleanly in six weeks. The drama is not always in the wound itself. It is in the aftermath.
Now. Did ChatGPT Write Her Book?
No. It did not. It gave Sarah the same information a medical historian would have given her over a two-hour consultation, except it took four minutes and cost nothing. The scene she went on to write, the one where her healer character packs yarrow into a soldier's wound while rain hammers the tent walls and the patient bites down on a leather strap, that was all Sarah. Her voice, her imagery, her pacing, her character choices.
ChatGPT told her what yarrow does. She decided what it meant.
The Conflation Problem
The hysteria in writing communities around AI is real, and some of it is justified. There are legitimate concerns about AI-generated content flooding the market, about publishers and readers being deceived, about the ethics of training data. Those are serious conversations worth having.
But those conversations have nothing to do with what Sarah did. Using AI to research medieval wound care is not the same thing as having AI write your novel. Treating them as equivalent is like saying a carpenter who uses a power drill has betrayed the craft of woodworking.
When you type a question into Google, you are using an algorithm built by a corporation to surface information from the web. When you type a question into ChatGPT, you are using a different kind of algorithm to synthesize and present information. The category is the same: you are using a tool to find out something you did not already know. The tool is more sophisticated. The research task is identical.
Your Job Has Not Changed
Here is what AI cannot do, no matter how sophisticated it gets: it cannot write your book. It can generate words, yes. But it cannot develop your characters from the inside out. It cannot make the structural decisions that give your story its shape. It cannot find the specific detail that makes your world feel lived-in rather than assembled. It cannot write a sentence that sounds like you, because it does not know what you sound like. Not really.
The craft is still yours. The voice is still yours. The story is still yours. What AI can do is fill in the research gaps quickly, so you spend less time in Wikipedia rabbit holes and more time actually writing.
The distinction that matters: Did AI write your prose? That is a problem. Did AI tell you how yarrow works so you could write the prose yourself? That is research. Those are not the same thing, and it is worth being clear-eyed about the difference.
Give It a Try
If you are writing historical fiction and you need to know how a trebuchet works, ask ChatGPT. If you are writing a thriller and you need to know how long it takes for a sedative to take effect, ask ChatGPT. If you are writing a contemporary drama and you need to understand the process of a specific legal proceeding, ask ChatGPT. Then go write the scene in your own voice, with your own characters, using the information as the foundation it is.
You are not betraying the craft. You are doing research, the same way writers have always done research, just faster and with fewer browser tabs.
Sarah, for what it is worth, finished her book. The wound scene is one of her best. She told me later that once she had the real information in front of her, the scene practically wrote itself. The healer's hands knew what to do, because Sarah finally knew what they should be doing.
That is what good research does. It gets out of the way and lets the story happen.
Hi, I'm Tim Hewitt
CEO of AuthorHub.io
I won my first writing scholarship in high school by beating more than 300 other writers in a nationwide contest, and that award gave me something money couldn't buy: the belief that my words mattered. Since then I've spent decades learning what makes writing work: how to find your voice, how to keep readers turning pages, marketing that really sells books, and how to spot the small craft problems that quietly sink good stories. This blog is where I share everything I've picked up along the way.
