There is a moment, new to medicine and now repeated in examination rooms across the country, when a physician opens the note for a visit that has just ended and finds it already written. Whenever it first happens to a given doctor, it tends to have the same shape. There is a lift of relief, which is genuine. Then comes something harder to name: the experience of reading one’s own name beneath prose one did not compose.
The title of this essay speaks of a first time, so I should say plainly what it will and will not do. It does not recount a particular morning. It is about a category of experience: the first time, in the long history of clinical writing, that the note arrives composed by something other than a person, and the first time, for each physician who meets it, that the ordinary relation between writing and authorship is turned inside out.
The relief deserves to be taken seriously. A finished paragraph looks finished. The headings are in place, the sentences join, and a conversation full of pauses, repetitions and half-sentences has acquired a beginning, a middle and a plan. Anyone who has spent evenings rebuilding visits from memory and scraps will understand the pleasure of that orderly page. I would not want to write about this change as though the pleasure were naive. Documentation has taken a great deal of time from physicians and, through them, from patients, and a tool that gives some of it back is not to be sneered at.
But imagine, for the length of this essay, a physician who is no one in particular, at the end of an afternoon, reading such a draft. The visit is recognisably there: the woman with the cough, the medicine she takes, the examination, the plan to review in a fortnight. The physician’s eye moves down the page with the pleasant sensation of having already done the work. Then it snags. There is a sentence in the history that she does not remember being said. It is not absurd; a patient with a cough might well say it. She cannot tell, from the page, whether this patient said it, whether she herself said it while thinking aloud, or whether it was supplied by a system that has read a great many coughs. And the page, in its calm typographical confidence, does not help her decide.
That snag is the subject of this essay. A machine-drafted note offers relief precisely because it looks like completed work. That is also what makes it a test of authorship. If the physician’s name goes beneath the page, the signature has to mean more than I recognise the general shape of this visit. It has to mean that the meaning survived her reading.
Hands That Were Not the Author’s #
Medicine has had ghostwriters before. For much of the twentieth century physicians dictated their notes and letters, to a stenographer and later into a machine; a transcriptionist typed them, and the physician signed. A phrase grew up around that arrangement and survives in places: dictated but not read, printed at the foot of a letter to warn its reader that the person whose name appeared there had not checked the words. Doctors have argued about the practice in the correspondence columns of the BMJ, and the Canadian Medical Protective Association, which advises Canadian physicians, cautioned in 2016 that an unreviewed entry risks incorrect information becoming part of a patient’s record indefinitely.
Later came speech recognition, which removed the typist and left the physician to correct the machine’s hearing himself. Its errors had a recognisable character: a homophone, a dropped negative, the name of a drug turned into an ordinary English word. They were often absurd, and absurdity is a kind of mercy, because it catches the eye. The drafts produced by today’s systems fail differently. Their errors, when they occur, are not misheard words. They are plausible sentences, grammatical and on topic, which is to say that they look exactly like the sentences around them. Each of these technologies changed who held the pen. This one also changes what a mistake looks like.
Literature knows the arrangement better than medicine does, and has thought about it longer. Milton was totally blind by the early 1650s, and Paradise Lost, published in 1667, was composed in his head and dictated to whoever was at hand: relatives, former pupils, friends, paid helpers. One early life of the poet records that he would compose verses in the night and, if his amanuensis came later than usual in the morning, complain that he wanted to be milked. Painters, above all in the nineteenth century, loved the scene of the blind poet dictating to his daughters, though the evidence for it is thinner than the paintings suggest; his nephew Edward Phillips recorded that two of the daughters were made to read aloud to him in languages they had not been taught to understand. Whatever the exact distribution of hands, nobody doubts whose poem it is. The words originated with Milton, and the duty of every scribe was fidelity to them.
What makes the case more than an anecdote is that Milton himself imagined the chain going one link further back. At the opening of the ninth book he describes his heavenly muse as a patroness
who deigns Her nightly visitation unimplored, And dictates to me slumb’ring, or inspires Easy my unpremeditated verse.
The poet who dictated to amanuenses presents himself as the amanuensis of a muse who dictates to him in his sleep. It is a conceit, and a humble one, and it reminds us that authorship has always been layered. What made the arrangement trustworthy was not that one person did everything. It was that each link knew its office: the muse inspires, the poet composes, the scribe records, and the poet, at least in principle, answers for the result.
Two centuries later, in the autumn of 1866, Fyodor Dostoevsky was in serious trouble. He had signed a contract with the publisher Fyodor Stellovsky that required a new novel by the first of November; if he failed, Stellovsky would acquire the right to publish his works without paying him for nine years. With weeks remaining and nothing written, Dostoevsky hired a young stenographer, Anna Grigoryevna Snitkina, recommended by the teacher of the shorthand course she had attended. He dictated The Gambler to her in under a month. She took his words down in shorthand, transcribed them at home, and brought the clean pages back for him to revise. The novel was finished in time, though Stellovsky made himself hard to find, and Dostoevsky is said to have lodged the manuscript with a police official in order to obtain a dated receipt: a document, that is, to certify the delivery of a document. The following February they married, and she went on to take down much of what he wrote for the rest of his life.
The detail I value in that story is the return of the pages. Snitkina’s shorthand was faithful, but faithfulness was not the end of the process. Dostoevsky read the transcription and changed it. If the old warning was dictated but not read, their arrangement was dictated and read, and the reading was where the author reasserted himself over the text.
The machine-drafted note is not dictation, and the difference is the heart of the matter. A transcriptionist’s duty is to reproduce what was said, and her work can be checked against the recording. A system that drafts a clinical note from a conversation does something else: it listens to two or more people, selects what seems relevant, discards the rest, reorders what remains into a conventional structure, and rephrases it in a professional register. That is synthesis, not transcription. Its fidelity cannot be checked against what was said, because it was never meant to reproduce what was said. It has to be checked against what was meant and what was established, and only the people in the room can know those things.
The Petition Writer #
Punjab has its own version of the arrangement, and it seems to me closer to the new situation than either Milton or Dostoevsky. Outside the district courts and revenue offices sit the ਅਰਜ਼ੀ-ਨਵੀਸ (arzī-navīs), the petition writers, whose name joins the Persian-derived words for a petition and a writer. The courts keep registers of licensed petition-writers, who give a person’s grievance the shape an office will accept. A farmer arrives with a dispute about a boundary or a delayed payment and tells it in his own words, which are full of history, grievance and digression. The writer listens, extracts what the office will care about, renders it in the formulas the office expects, and hands it back. The facts belong to the petitioner, who signs, or presses a thumb to the page; the form and much of the wording belong to the writer. Many petitions in the English of Indian courts still open with an old formula, Most respectfully sheweth, and proceed through numbered paragraphs that make a widow’s claim to a field and a shopkeeper’s complaint about a neighbour sound remarkably alike. The register falsifies nothing; it is what allows an office to process a thousand grievances. But it is also the reason a petitioner might hear his own petition read aloud and not quite recognise it.
Everyone involved understands whose petition it is. Everyone also understands that the petitioner may not be able to check every phrase. Punjabi has an unkind colloquial term, ਅੰਗੂਠਾ ਛਾਪ (angūṭhā chhāp), thumb-stamp, for a person who signs with a thumb because he cannot read. The insult is unfair to the petitioner, who is usually intelligent and often shrewd. But it names something real about the transaction. A petition can be read back aloud before it is signed, and when it is not, the thumbprint vouches for words its owner never heard.
The machine-written note rearranges these roles in a way I find genuinely new. With dictation, the words were the physician’s, and the typist’s duty was fidelity to them. With the petition writer, the facts were the client’s and the form was the writer’s. In the ambient arrangement, two people speak, something else chooses the words, and the physician becomes the first reader of his own note. Authorship used to begin with composing. For this kind of writing it now begins with recognition.
That shift carries a specific danger, and the petition writer’s register names it. The physician’s risk is not that he cannot read. It is that he is a literate signer with too little time to read, whose signature then does exactly what a thumbprint does on an unread petition: it vouches, in his name, for a text he has not really heard.
Auctor #
The word itself is suggestive. Author comes, through Latin auctor, from augere, to increase or make grow, the verb that also gives English augment. In classical Latin an auctor was not only an originator. The word could name a founder, a source, an adviser, and also a guarantor: the person who vouches for something and stands behind it, whose auctoritas lends it weight. In one of its oldest senses, authority was the capacity to warrant.
The American Medical Association adopted augmented intelligence as its preferred term in 2018, to stress that such tools should enhance clinical judgment, not stand in for it. If the etymology is any guide, an author and an augmenting tool are in related trades: both make something larger. What ought to grow, in a clinical note, is the visit itself, made legible to the next reader. A note that has grown away from the visit has been augmented in the wrong direction.
The Latin also explains why the signature matters more than the composition. A note composed by a machine can still have an author in the old sense, someone who warrants it. That is what a signature is: not a claim to have chosen every word, but a guarantee that the words can be trusted to mean what they say. A guarantor who has not inspected what he guarantees is not an author in either sense. He is a name.
Punjabi, borrowing from Persian, calls a signature ਦਸਤਖ਼ਤ (dastakhat), from the words for hand and writing. The etymology is plain, and it has become newly pointed. In a note drafted by a machine, the signature may be the only part of the page that was, in any sense, written by the physician’s hand, and it is the part that every later reader will rely on most.
The Made Thing #
The most famous story in English about a maker and the thing he made is, among other things, a story about authorship. Mary Shelley began Frankenstein in the summer of 1816, at a villa near Geneva, when Lord Byron proposed that each member of his small party write a ghost story; it was published anonymously in 1818. In the introduction she added to the revised edition of 1831, she reflected on where the book had come from, and wrote a sentence that could serve as a description of a drafted clinical note: “Invention, it must be humbly admitted, does not consist in creating out of void, but out of chaos.” The materials, she went on, must first be afforded; invention can give form to dark, shapeless substances, but cannot bring the substance itself into being.
That is exactly what such a system does with a visit. It gives form to chaos: the pauses, the interruptions, the patient who answers the question before last, the physician thinking aloud. It cannot create the substance. But it can produce form so convincing that a reader assumes the substance was there, and when a gap in the conversation meets a template’s expectations, the form will sometimes fill it.
Victor Frankenstein’s failure is often described as hubris, the overreach of making life. Read again, the novel seems to me more interested in a different failure. Victor does not stay with what he has made. At the moment the creature opens its eyes he rushes from the room, and for the rest of the book he refuses, until far too late, the responsibility that making it entailed. The catastrophe follows less from the act of creation than from the abandonment after it.
There is a passage in the creature’s own narrative that I find hard to read. Having learned to read, he discovers in the pocket of a coat he took from the laboratory his maker’s journal of the months before his creation. He reads the whole record of his making, and with it the description of himself, written in a language that conveyed his maker’s disgust more vividly than any fact about him. It is one of the great scenes in English fiction of a subject reading his own record. The note that describes a patient will one day be read by her, too; that is a subject for another essay. Here what matters is the other direction. The creature read the document and found the maker in it.
And the made thing took its maker’s name. In popular usage the creature, who has none in the novel, has long been called Frankenstein. I can think of no better emblem for a machine-drafted note. It will carry the physician’s name whether or not the physician ever really looked at it, and whatever it says will be understood by every later reader as his.
The book’s own history adds a last turn. When Frankenstein appeared anonymously, at least one eminent reviewer, Walter Scott, assumed that its author was Percy Shelley. Scholars working from the surviving manuscript notebooks have since shown Percy’s hand throughout the draft, in revisions, substitutions and additions, and have argued about how much they changed it. Nobody serious now doubts that the novel is Mary Shelley’s. Help, even extensive help, does not dissolve authorship. What makes the book hers is that she governed it: the conception, the judgment about what stayed and what went, and the willingness, in 1831, to claim it publicly and bid her “hideous progeny go forth and prosper.” Authorship, in that sense, is the act of standing behind a text, and it is available to someone who did not choose every word.
A Conversation Becomes a Document #
What does standing behind a drafted note require? It helps first to see the stages by which talk becomes record, because each has its own way of losing things.
Imagine a fragment of a visit, invented for the purpose. The patient says: It’s been on and off, mostly when I climb the stairs up to the flat, the chest goes tight, like a band, but then last Sunday it happened sitting down, which is why the family made me come. The physician asks a few questions; the patient answers some of them twice and one of them not at all; a telephone rings; they return to the subject later from another angle.
The first stage is capture: who said what, in what order, with the hesitations intact. Nothing is lost yet except gesture and tone, which is already a great deal. The second is selection. Most of what was said is discarded, because most of what is said in any conversation is scaffolding. The third is organisation. The surviving material is sorted into the headings of the conventional note, a structure that descends from Lawrence Weed’s problem-oriented medical record of the late 1960s: what the patient reports, what the clinician observes, what the clinician concludes, what will be done. Those headings are not neutral containers. They embody a theory of the visit in which the patient supplies data and the clinician supplies meaning, and a remark that fits no heading tends to vanish or to be forced into the nearest one. The fourth stage is translation into the professional register. The chest goes tight, like a band becomes chest tightness; when I climb the stairs becomes exertional; last Sunday it happened sitting down becomes, if it survives, one episode at rest. And which is why the family made me come, which explains the timing of the visit and the worry behind it, may not survive at all.
A physician writing his own note performs every one of these stages too. The difference is that he performs them with the conversation still in his head, so that each compression is checked, however imperfectly, against memory as it is made. When a system performs them, the physician meets only the result. The compression has already happened. He must work backwards from product to process, and he must notice what is missing from a page that gives no sign anything was ever there.
A Possibility Becomes a Decision #
The difference between a faithful draft and an unfaithful one can be very small, so it helps to be concrete about particular sentences too. Imagine a few more exchanges, invented as language and not taken from any record, and the forms a compressed summary might give them.
The physician says, If the pain is no better in a couple of weeks, we could think about a scan. The draft reads: Plan: imaging in two weeks. The grammar is impeccable. A conditional has become a schedule, and a thought spoken aloud has become an order that someone may now expect to happen. Certainty has been added.
The patient says, I wonder if it’s the stress at work. The draft reads: Symptoms likely related to occupational stress. What has moved here is ownership. A patient’s speculation has become a clinician’s assessment, with a probability attached that nobody in the room expressed. Attribution has changed.
The patient says she has been taking a medicine most days, except when an early shift makes the morning difficult. The draft reads: Taking medication as prescribed. The sentence is ordinary. It is also different. The missing difficulty may be exactly what the next conversation needs to address, and restoring it is not a stylistic preference; it puts back information that could matter to the plan. Something has been omitted.
The patient says the pain began after a fall last month, though she had had twinges before. The draft reads: Pain since fall one month ago. The earlier twinges, which might change what the fall means, have vanished. Chronology has been simplified.
Asked about fever, the patient says, I don’t think so. Maybe once, last week, I felt hot. The draft reads: Denies fever. A hesitant answer with a qualification has become a clean negative. Negation has been hardened.
Neither sentence in any of these pairs is absurd, which is exactly why each is easy to pass over. A useful test of the draft is therefore not whether the physician recognises the visit. Recognition can be surprisingly forgiving. A few familiar symptoms, the right medication name and the general direction of the plan may be enough to make a page feel like an account of the conversation. But resemblance is a lower standard than accuracy.
Review, in this arrangement, is not proofreading. It means noticing a sentence that sounds reasonable and says more, or less, than the encounter established. That is a different mental act from composing. A writer starts from what he means and searches for words; a reviewer starts from words and must reconstruct whether he meant them, and anyone who has reread his own work knows how readily the eye supplies what it expects to see. Researchers who study people working alongside automated systems have a name for one part of this, automation bias: the tendency to accept a machine’s output with less scrutiny than the same claim would receive from a person, and to miss what the machine has failed to flag. Nobody need be careless for it to happen. Attention simply behaves this way when it has been told, implicitly, that the work is done, and a finished-looking page makes that announcement on every line. The most persuasive departures are the least conspicuous: a qualification removed, a speaker changed, a plan made firmer than the decision was. An ethics analysis by Sara Gerke and David Simon in the AMA Journal of Ethics in 2025 noted that such tools can introduce content nobody said, and warned against letting the burden of catching it drift onto patients. Patients can now read most of their notes, and some of them will find such sentences. But a patient who finds a sentence she never said will reasonably assume that her doctor wrote it, and she should not be made the system’s quality control. That burden sits with the person whose name is on the note.
Borges and I #
There is another change, quieter than error, that a drafted note brings with it, and it concerns voice. A physician’s own notes, however hurried, carry traces of a particular person: a habitual way of recording uncertainty, a preference for the patient’s own words in quotation marks, a phrase that captured exactly what this woman was afraid of. A drafted note tends to replace those traces with a standardised clinical register, fluent and interchangeable. It also speaks in the first person. I discussed the risks and benefits with the patient, who verbalised understanding. The physician reads a sentence beginning with I that he did not write, describing an act in language he would never have chosen.
Jorge Luis Borges wrote a page about this feeling long before it had anything to do with medicine. In “Borges and I”, collected in El hacedor in 1960, the narrator describes the other, public Borges, to whom things happen and whose name turns up on a list of professors or in a biographical dictionary. The narrator lets that other self go on producing literature, and finds that he recognises himself less in Borges’s books than in many others. The piece ends with a sentence that, in my rendering of the Spanish, reads: “I do not know which of us is writing this page.”
The drafted note creates a clinical Borges: the physician of the record, who always counselled extensively, whose patients always verbalised understanding, and who speaks in a voice assembled from a great many other notes. No physician needs the machine’s sentences to sound like him. They will not, and much of the time it hardly matters. Style is a preference, and preferences are cheap to give up.
But the distinction between style and substance is not always where it seems. Sometimes the personal phrasing was carrying information. She is worried she won’t be able to lift her grandson says something that patient concerned about functional limitations does not, and the difference is not decoration; it tells the next clinician what this person cares about and therefore what a good outcome would look like. Standardisation strips out exactly that kind of particular. So the reviewer has two jobs that should not be confused. One is to let go of style without regret. The other is to notice when the loss of a particular phrase has taken patient-specific meaning with it, and to put that meaning back.
Borges could end his page in uncertainty about which of them was writing it. That uncertainty was the point of the piece, and literature can afford it. A clinical signature cannot. The physician may freely admit that something else wrote the page. What he must not be uncertain about is whether he read it.
Reading Before Signing #
What would adequate reading look like? I can describe the routine I would want any signer of such a note to follow, offered as an illustration that any practice would need to test and adapt, not as a standard of care.
Read the history for attribution and time. Is every statement attributed to the patient something she said? Has anything she merely wondered become something she reported? Are the dates and sequences right, and have any qualifications fallen away?
Read the examination for what was actually examined. A drafted note may fill in normal findings for parts of the body that were never looked at, because templates expect them, and a normal finding recorded but not observed is a particularly dangerous kind of error: it closes a question no one asked.
Read the assessment for certainty. Does each conclusion carry the degree of confidence that was actually expressed? Are possibilities still possibilities?
Read the plan for agency and condition. Who will do what, by when, and under what circumstances? Have conditionals stayed conditional?
Read the medications line by line, for names, doses, changes and the reasons for them, and check any phrase such as as prescribed against what was actually said about taking them.
Read the follow-up for what was agreed, not what is customary.
There is one further check, borrowed from the petition writer’s better habit, that seems to me worth trying where time allows: reading the essentials back aloud before the patient leaves. So the plan is that you’ll keep a note of when the pain comes, and if it isn’t better in two weeks we’ll talk about a scan. Is that what you understood? A sentence like that is no part of the record. Yet it settles in the room, with both parties present, the very points a draft is most likely to harden or blur, and it gives the person described a chance to say that isn’t quite what I meant while the conversation can still be changed.
Stated like that, the routine sounds obvious, and it should. The difficulty lies less in knowing what to check than in having the conditions in which checking is possible.
Here the strongest objection to everything I have written deserves its due. Demanding careful review is necessary, but it can become an empty instruction. If a physician’s schedule is built on the assumption that drafted notes need only a glance, the time saved by the draft will be spent on more visits, and the careful reading this essay recommends will happen, if it happens at all, in the evening, unpaid and tired. If the interface presents a finished-looking page with no way to see which sentence came from which moment of the conversation, the reviewer must rely on memory, which is the very faculty the tool was meant to relieve. In those conditions, telling the individual physician to read carefully is like telling the petitioner to read his petition when no one has offered to read it aloud. The responsibility is real; the means of discharging it have been withheld.
So the signature depends on things the signer does not entirely control. Adequate time for review has to be built into the work, not carved from the edges of it. The physician needs full authority to correct the draft, including the authority to delete sentences that sound plausible and cannot be confirmed. The page needs traceability: a way to see, for any sentence, where in the conversation it came from, so that review can be a check rather than an act of recollection. And there needs to be an honest account of uncertainty, both in the note and about the tool.
That last point extends beyond any one note. When a physician corrects a draft, it is tempting to imagine that the system learns from the correction, as a human trainee would. That should not be assumed. Whether edits are captured, analysed or used to change a tool depends on how the product is built and governed, which the signer often cannot see. A correction made in one note fixes one note. If the same kind of error recurs, a qualification dropped here, a speculation promoted there, the pattern will only change if someone collects it, reports it and treats it as information about the tool: something to consider when systems are chosen, when clinicians are trained to use them and when their performance is monitored. Individual vigilance catches individual errors. Only an institution can learn from them.
The Name Under the Paragraph #
None of this is an argument against the tools. I would happily hand over repetition, and the time many physicians spend reconstructing visits at the end of a day can be put to far better use. Composing a note was never sacred, and nobody should romanticise it; the handwritten notes of the past were often illegible, the dictated ones often unread. But I would want enough attention left, after the relief, to examine whatever had been produced in my name. Assistance should make it easier to stand behind the record, not harder to see what one is standing behind.
Return, then, to the imagined physician and her snagged sentence. Whether the machine wrote well is beside the point. The question she faces is whether she can warrant what it wrote, in the old sense of auctor. If she can trace the sentence to the conversation, she can keep it. If she cannot, and her memory does not supply it, the honest course is to remove it, however plausible it sounds. A plausible sentence that cannot be confirmed is exactly the kind of sentence that will be believed by everyone who reads it later, because it has her name beneath it.
At the foot of a finished note the record prints a line declaring that it was electronically signed, with a name and a time. That line has always meant the same thing, whether the words above it were written by hand, dictated to a typist or drafted by a machine: this is what happened, as far as I can tell. Milton’s amanuenses took down his words; Snitkina returned Dostoevsky’s pages for him to change; the petitioner’s thumb vouches for what the writer has made of his grievance; Mary Shelley sent out a book that other hands had touched and called it hers. In each case the work became the author’s at the moment someone stood behind it and could say why.
Sometimes the most important edit before that line is not a correction of fact at all. It is a single word put back, the one the draft removed because it made the sentence less tidy and the plan less firm. Could.
Evidence and Editorial Note #
Documented findings. The essay relies on the author’s public biography: a family physician practising in California since 2010, with Fresno as home, whose first language is Punjabi. The phrase “dictated but not read”, its use as a warning that a text has not been checked by its signatory, and critical BMJ correspondence about the practice (2001): Wikipedia, “Dictated but not read”, and the letters it cites. The 2016 caution that unreviewed dictated or speech-recognition entries risk incorrect information remaining in the record indefinitely: Canadian Medical Protective Association, “’Dictated but not read’: Unreviewed clinical record entries may pose risks” (June 2016, reviewed 2021). Licensed petition-writers registered by district courts in Punjab: Punjab and Haryana High Court Rules and Orders, Volume I, Chapter 17. Author from Latin auctor, from augere, “to increase”; augment from the same verb: Online Etymology Dictionary; the senses of auctor as founder, source and guarantor are standard Latin lexicography. The American Medical Association’s 2018 adoption of “augmented intelligence” as its preferred term: AMA policy as reported by Becker’s Hospital Review. Ambient tools’ capacity for invented content, and the caution against placing the onus of error-catching on patients: Gerke and Simon, AMA Journal of Ethics, November 2025. Literary and historical sources: Milton’s blindness by the early 1650s, the publication of Paradise Lost in 1667, its dictation to various amanuenses, the anecdote of wanting to be “milked” from an early anonymous life, and Edward Phillips’s account of the daughters reading aloud in languages they did not understand, from the early lives as commonly reported; Paradise Lost, Book IX, lines 21–24, quoted from the public-domain text in modernised spelling. Dostoevsky’s 1866 contract with Fyodor Stellovsky, the November 1 deadline and nine-year penalty, the dictation of The Gambler to the stenographer Anna Grigoryevna Snitkina (recommended by her shorthand teacher), her transcription and his revision, the delivery of the manuscript to a police official for a dated receipt (reported, hence “is said to”), and their marriage in February 1867: standard biographical accounts, including her memoirs. Lawrence Weed’s problem-oriented medical record (New England Journal of Medicine, 1968) as the source of the conventional note structure; automation bias as a recognised human-factors concept; the formula “Most respectfully sheweth” in Indian court petitions; ਦਸਤਖ਼ਤ from Persian words for hand and writing: standard reference knowledge. Mary Shelley, Frankenstein (1818; revised edition 1831), the 1816 ghost-story origin, the 1831 introduction (two short public-domain quotations), Victor’s flight, the creature’s discovery of the journal, the popular transfer of the maker’s name, Walter Scott’s 1818 review attributing the book to Percy Shelley, and the manuscript notebooks’ evidence of Percy Shelley’s revisions: the novel’s text and standard scholarship. Jorge Luis Borges, “Borges y yo”, in El hacedor (1960), paraphrased, with the closing line in the author’s own rendering from the Spanish. The Punjabi terms ਅਰਜ਼ੀ-ਨਵੀਸ and ਅੰਗੂਠਾ ਛਾਪ are given with their ordinary meanings. Vendor representations. None; no product, vendor or health system is named or assessed.
Recollection and tradition. No first encounter with a machine-written note, and no specific note, product, employer, colleague or patient, is described; the essay does not assert that the author has used or signed such a note. The physician reading a draft with a snagged sentence is hypothetical. The wording comparisons are invented language examples and do not describe a real patient or record. The review routine is an illustration, not a description of any institution’s policy.
Analysis and forecast. The account of authorship as recognition and warranty, the readings of Milton, Dostoevsky, Mary Shelley and Borges as they bear on clinical notes, the review routine, and the author’s views on organisational conditions for meaningful signing are personal interpretation as of November 6, 2026. Nothing here is documentation, legal or compliance guidance. Independence. This essay is not sponsored by any institution, company, health system or technology developer. It is personal reflection, not medical advice; readers seeking care should contact their own physician’s office.
About the Author #
Kanwar Partap Singh Gill, MD, traces his family roots to Khadur Sahib in Punjab’s Majha region and grew up in Amritsar. He attended Spring Dale School and completed pre-medical studies at Khalsa College in 1994–1996. He earned his MBBS from Sri Guru Ram Das Institute of Medical Sciences and Research in Amritsar in 2002.
He completed PGY-1 at the University of Arkansas for Medical Sciences in Little Rock in 2007–2008, and PGY-2 and PGY-3 with UCSF Fresno at Community Regional Medical Center from 2008 through 2010. He has practiced medicine in California since 2010.
His career has crossed four distinct information eras: predominantly paper-based medicine, early and fragmented electronic records, the integrated-EHR era, and the emerging age of generative and agentic artificial intelligence.