The AI tag is only used when most of the text was created with AI. Authors who use AI to create illustrations or small parts of a story shouldn't use the AI tag.
The AI tag is only used when most of the text was created with AI. Authors who use AI to create illustrations or small parts of a story shouldn't use the AI tag.
or small parts of a story shouldn't use the AI tag.
Disagree with you there. If you use AI to help you write, you are using AI. The way I see it, if you get help from someone on a test paper, even if it's only for two questions out of a hundred, it's still cheating. You have not done the test yourself. Using AI to create parts of your story, even if it's only "small parts", is still collaboration with an 'other'. It matters not if the 'other' is carbon based or silicon based.
Authors who use AI to create illustrations or small parts of a story shouldn't use the AI tag.
The definition of the AI Generated tag is:
"Text was generated mostly by an AI."
(bold mine.)
I would suggest that if an author wants to use AI to create small parts of a story, follow the following steps:
1) Have an AI write a scene;
2) Instead of copying and pasting what the AI wrote, take the portion written by an AI and rewrite it in their own words.
In other words, I see a legitimate use of AI being a tool to overcome a writer's block.
In other words, I see a legitimate use of AI being a tool to overcome a writer's block.
As an experiment, I just asked Grok to create a sequence of cases for an Emergency Department doctor. It did a credible job, and such a list could be used as a prompt to give me ideas.
I don't see this as different from watching, say ER or Emergency or Grey's Anatomy or The Pitt for ideas.
My point is, 'using AI to assist in writing' is different from 'AI writing', at least as I see it.
My point is, 'using AI to assist in writing' is different from 'AI writing', at least as I see it.
Uh oh, you've started down the slippery slope.
AJ
Uh oh, you've started down the slippery slope.
Did you miss the 'experiment' comment? I reported asking Grok to actually write a chapter several weeks ago, and it was absolute sh-te.
This is done out of curiosity, not as a plan to make use of it.
I will repeat the question, though — how different would be seeking ideas from AI versus watching TV or movie medical, legal, or financial dramas?
how different would be seeking ideas from AI versus watching TV or movie medical, legal, or financial dramas?
Probably very similar and equally inaccurate if the AI was trained using stolen books rather than real life ER data.
AJ
Probably very similar and equally inaccurate if the AI was trained using stolen books rather than real life ER data.
I didn't ask it to describe the scenes, just for a sequence of presenting patients. It came up with something credible, as I would expect.
When I did ask it to write a chapter, it did a terrible job, one that anyone who read my stories would instantly recognize as sh-te AI content.
It came up with something credible, as I would expect.
I would question the word 'credible'. 'Entertaining' perhaps, but with little semblance to reality.
AJ
I would question the word 'credible'. 'Entertaining' perhaps, but with little semblance to reality.
It was simply a suggested list of cases seen by a doctor in the Emergency Department over the course of a shift. That's not difficult, as many such lists exist, as do statistical studies.
Perfectly credible.
That's not difficult, as many such lists exist, as do statistical studies.
That's different from taking lists from entertainment sources.
AJ
That's different from taking lists from entertainment sources.
Depends on the source and how you filter (e.g., my prompt expressly limited outlier cases). I can do the same thing mentally when watching TV. The AI can filter the same way (per the prompt).
Judge for yourself…
Prompt:
Produce a statistically likely list of 20 cases a doctor in an Emergency Department in a small city might see during a shift, with most cases in the first standard deviation, but one case outside the first standard deviation of likelihood. Include a working diagnosis.
Response:
1. 29-year-old female – lower abdominal pain, dysuria, and urinary frequency for 2 days. Working diagnosis: uncomplicated UTI.
2. 41-year-old male – right ankle pain and swelling after twisting it on uneven ground at a construction site. Working diagnosis: lateral ankle sprain.
3. 7-year-old child – fever (101°F), cough, and rhinorrhea for 4 days. Working diagnosis: viral upper respiratory infection.
4. 52-year-old male – acute midline low back pain after lifting heavy boxes at work. Working diagnosis: musculoskeletal lumbar strain.
5. 38-year-old female – crampy abdominal pain, nausea, vomiting, and diarrhea after a family picnic. Working diagnosis: viral gastroenteritis.
6. 65-year-old male – worsening shortness of breath and wheezing with known COPD history. Working diagnosis: COPD exacerbation.
7. 25-year-old male – 3 cm laceration to forearm from a utility knife at home. Working diagnosis: simple laceration.
8. 44-year-old female – throbbing unilateral headache with nausea and photophobia, history of migraines. Working diagnosis: acute migraine.
9. 3-year-old male – fever, vomiting, and mild dehydration after a day of daycare exposure. Working diagnosis: viral syndrome.
10. 30-year-old female – diffuse hives and mild lip swelling 30 minutes after eating peanuts (known allergy). Working diagnosis: anaphylaxis (mild).
11. 48-year-old male – neck pain and stiffness after rear-end MVC on a rural highway. Working diagnosis: whiplash/soft-tissue injury.
12. 42-year-old male – sudden left flank pain radiating to groin with microscopic hematuria. Working diagnosis: ureteral nephrolithiasis.
13. 68-year-old male – dizziness and near-syncope while standing, history of hypertension. Working diagnosis: orthostatic hypotension or vasovagal.
14. 26-year-old female – palpitations and anxiety after a stressful day, no chest pain. Working diagnosis: panic attack.
15. 19-year-old male – sore throat, fever, and tonsillar exudate. Working diagnosis: streptococcal pharyngitis.
16. 49-year-old female – facial swelling and severe tooth pain for 3 days. Working diagnosis: dental abscess.
17. 58-year-old male – red, warm, swollen lower leg with low-grade fever. Working diagnosis: cellulitis.
18. 12-year-old male – minor head bump from fall off bike, brief loss of consciousness. Working diagnosis: concussion evaluation.
19. 55-year-old female – substernal chest pain relieved by antacids, history of GERD. Working diagnosis: non-cardiac chest pain (GERD/musculoskeletal).
20. Outlier (outside first standard deviation of likelihood) – 59-year-old male – sudden-onset severe tearing chest and interscapular back pain radiating from a hypertensive crisis (BP 210/110). Working diagnosis: acute aortic dissection (Type A suspected). This presentation is a classic high-mortality zebra: while nonspecific chest/back pain is among the top 3–5 reasons for ED visits nationally, aortic dissection occurs in only a tiny fraction of such cases (≈1 in 1,000 chest-pain presentations and far rarer overall).
That's significantly different from what UK A&E doctors are likely to see, especially on nights that aren't 'school nights' (in the going to work the next day sense. A large proportion of their manpower is wasted on drunks and druggies.
AJ
A large proportion of their manpower is wasted on drunks and druggies.
Why the prompt expressly stated 'small city'! I didn't want 10 gunshot wounds and 10 drunks! 😜
The List of possible admissions to a small city ER seems Reasonable to me; albeit my knowledge is from the perspective of an EMT decades ago, and of Military (and Naval) TMCs/ERs into c.2016 or so. Also, several friends who are EMTs, Paramedics, Nurses, or Doctors.
(I would have expected 1 or 2 URI and 1 or 2 drug/alcohol exacerbated situations; however, that may be because I live in the PNW (Pacific Northwest) (Military/Naval bases) also having those issues.
Region has a significant influence upon "typical" admissions to an ER.
After my retirement from the US Army, where my duties resulted in knowledge of what occurred in the TMC (Troop Medical Clinic)[similar to neighborhood clinics], and the ER; I now volunteer to transport fellow veterans to the VA, including the ER after Clinic hours. Then I speak with people in the field about "how common" particular issues are. Also, about adding items to my various medical kits (vehicle, camping, range, "Bug Out" EDC, home; each having different priorities).
(I would have expected 1 or 2 URI and 1 or 2 drug/alcohol exacerbated situations; however, that may be because I live in the PNW (Pacific Northwest) (Military/Naval bases) also having those issues.
The posted prompt didn't include the date — 1992 — which would change the types/kinds of things seen compared to 2026.
I'm almost afraid to ask it for Cook County (Stroger Hospital) in 2026!
Such lists are interesting. I would Not "copy" such a list, but it is a reasonable primer, from which Fictious, but Plausible visits to an ER may be crafted for a story.
Because the demographics of the USA are considerably different now than in 1942, I used several sources for my stories about WWII. NYC is the most populous city in the USA. I didn't know that some 4% to 7% of All personnel who served in the armed forces of the USA came from NYC (including the outter boroughs) [somewhat skewed by men who came to NYC to enlist, or for a civilian job and then enlisted/were drafted]. Verified by research of Service Numbers, which were issued based upon Place of Residence (or, less often, place of Enlistment).
Tropes common to most USA movies about WWII of a Guy from NYC make sense if you consider that a Pltoon of 40 Men would likely have 1 to 3 men from NYC; statistically...
Of course, units of the NYNG (New York National Guard, in particular units based in or near NYC, would likely have a higher percentage (by early 1944, most National Guard units were composed of 50% Draftees!
Chicago was the only other city nearly as large. A
Because the demographics of the USA are considerably different now than in 1942, I used several sources for my stories about WWII. NYC is the most populous city in the USA. I didn't know that some 4% to 7% of All personnel who served in the armed forces of the USA came from NYC (including the outter boroughs) [somewhat skewed by men who came to NYC to enlist, or for a civilian job and then enlisted/were drafted]. Verified by research of Service Numbers, which were issued based upon Place of Residence (or, less often, place of Enlistment).
In the 1940 census, about 5.6% of the US population lived in NYC (NYC population was 7,454,995 while the US population was 132,164,569)
I didn't know that some 4% to 7% of All personnel who served in the armed forces of the USA came from NYC
That included my dad, who was in the Naval Reserve, living in NYC at the time of Pearl Harbor. He'd joined the reserves immediately after graduating from Dartmouth (1939) and went to work as an Electrical Engineer. He was called to active duty and served as a Chief Petty Officer for the duration, usually assigned as the Chief Radio Man on destroyers running convoys in the Atlantic.
Actually, NYC is the largest, LA is second in population, while Chicago is third (the last I checked), but with so many now actively fleeing the country for better/safer options overseas, I suspect those numbers are still in flux. However, those numbers were consistent for most of the twentieth and twenty-first century, so I think you're safe.
In a sinking boat, sometimes the rats are the last to flee.
In a sinking boat, sometimes the rats are the last to flee.
My local convenience store has a rat problem. It's now closed for deratification. I expect it's the Monty Python generation: they like their convenience stores with not so much rat in them.
AJ
Actually, NYC is the largest, LA is second in population, while Chicago is third (the last I checked), but with so many now actively fleeing the country for better/safer options overseas,
Those are three of the top places people are fleeing from. But not overseas. To other parts of the U.S.
My point is, 'using AI to assist in writing' is different from 'AI writing', at least as I see it.
Uh oh, you've started down the slippery slope.
I don't believe that's a slippery slope. I think it makes sense.
Asking AI for another word for a word is no different than using a thesaurus.
Asking AI for a metaphor or simile for something is sort of the same as the previous one.
Asking AI for ideas where something could trigger something else happening is no different than asking someone in your writer's group that question.
AI isn't writing your story in those cases.
Where would you set the limit on the AI contribution to your story, given that you're using its words and plot?
AJ
Where would you set the limit on the AI contribution to your story, given that you're using its words and plot?
Writing it — like the whole thing or copying and pasting a section.
And I didn't mean for the AI to write the plot. More like asking it a question like: What are some reasons a priest might get blackmailed?
Still, you need to be careful. Like ultra-processed food, AIs are designed to be incredibly more-ish.
AJ
That's fair, as long as you don't include it in your writing (i.e. separate the automated background research from the actual written story). Duh!
"a tool to overcome a writer's block"? You mean, having no clue how to write yourself out of the corner you painted yourself into (a fairy common problem)?
As was noted earlier, you either know how to write or you don't, and thus need AI to write what you're unable to write yourself. That is the very definition if cheating, and it's much greater than just "one or two" answers on a test, as it's literally what you're incapable of finishing on your own.