Four hundred and one years ago Francis Bacon wrote that “Reading maketh a full man, conference a ready man, and writing an exact man.” Today, scrolling copy alone and having AI generate our writing appears set to prove the effects of the opposite behaviors.
One of the pleasures of writing something, of being clear in words that others will read, is finding out what you are actually thinking. When people use AI for their writing they corrode the quality of their thought.
We’d talk about this more, but too many of us have leaned into AI to form our thoughts effectively.
The Bacon quote is exactly right — and it's exactly the spirit in which I try to use AI: "Conference maketh a ready man." My best AI interactions are conference, not dictation. The AI offers a framing. I push back. It pushes back. Something neither of us started with emerges. The discipline of making my thinking precise enough to evaluate, extend, or reject someone else's contribution is at least as demanding as writing alone. Maybe more, because I'm developing my own position while engaging with another's — even if (or maybe especially if) the other in this case is a non-human intelligence.
The corrosion you're describing is real — but it's a symptom of how most people use AI, not of AI itself. The person who prompts "write me a memo" and sends whatever comes back has skipped the exactness Bacon was talking about. Their thinking never got precise. But that's a choice, not an inevitability.
These aren't AI writing for you. They're AI creating structured spaces where your thinking has to become exact — in Bacon's sense. The tool can corrode thought or discipline it. The difference is whether you're using it to avoid thinking or to think harder than you could alone.
Bacon would recognize the process, I think, even if the medium would surprise him.
"One of the pleasures of writing something, of being clear in words that others will read, is finding out what you are actually thinking. When people use AI for their writing they corrode the quality of their thought." I find the opposite. Note your own words " to form our thoughts effectively." Isn't that the purpose of writing?
I liked this essay, but the author engages in hyperbole when she states that language creates reality. Language emerges from reality; it is our (often feeble and inadequate) response to reality. But the underlying “ground of being” is not a word or even “The Word.” We humans congratulate ourselves on our clever use of words by making such claims, but the entire realm of nonverbal conduct, which often (always?) has more salience gives the lie to this categorical mistake.
I share some of these misgivings. The boss who pastes a ChatGPT congratulations email without reading it is being inauthentic — but not because AI was involved. They'd be equally inauthentic with a Hallmark card or a copied-and-pasted message from last year. The inauthenticity is in the absence of thought, not the presence of a tool.
But the essay defines authenticity as a specific process — the writer alone choosing each word — and treats any other process as inauthentic by definition. A collage artist doesn't create each image fragment from scratch. A photographer doesn't paint each pixel. Nobody calls these media inauthentic because the artist operates at the level of composition and vision rather than individual rendering. AI-assisted writing is a different medium that operates at a different level. The authenticity lives in the thinking, the judgment, the perspective — not in whether one human selected every word.
More importantly, the essay treats all writing as expression. Most writing isn't. It's functional. A single mother disputing a medical bill doesn't need authentic self-expression — she needs clear, legally informed language that protects her rights. Before AI, her options were write it herself in language that might be unclear or legally naive, or pay a lawyer she can't afford. AI gives her a voice she never had in a domain where her natural voice was a disadvantage. That's not inauthenticity. That's access.
A student in Ghana can ask AI to translate an English-language article into Twi and then engage with the ideas in their strongest language. The engagement — their questions, their connections to their own context — is deeply authentic. The translation removed a barrier, not an authenticity.
The hand-wringing about AI and writing comes disproportionately from people whose specific relationship to writing is threatened — journalists and creative writers who experience writing as identity. That's real and worth taking seriously. But most people experience writing as obligation: the email they have to send, the form they have to complete, the application they have to get right in a language that isn't their strongest. For them, AI isn't undoing authenticity. It's removing an obstacle between their intent and their expression. The intent was always there. The craft wasn't. Now it's available.
The real question isn't whether AI-assisted writing can be authentic. It's whether the person using AI bothered to think. That question has the same answer it's always had, with or without AI.
The entire media economy rests on a single premise: persuading people to part with their money for things of dubious or deceptive value. Advertising.
And the entirety of public discourse has decayed into a parallel enterprise: persuading people to vote for things of similarly dubious or deceptive value. Politics.
Before AI arrived, we had reached a collective conclusion that we could no longer trust anything anyone writes or says.
Enter AI.
This article insists that AI-generated content will destroy trust and authenticity. But how do you destroy what has already been reduced to rubble?
Trust has always been a personal attribute. It belongs to the person, not the tool. If society wants to recover it, we must relearn how to judge character — honesty, clarity, logic, common sense — rather than credentials, celebrity, or the sheer volume of one’s outrage.
And here is the irony: precisely because AI is inhuman, it may help strip away the human pretenses that have corrupted advertising and politics. If it exposes the hollow performances, the manipulations, the failed character behind the message, then authenticity might again become visible — not as branding, but as a personal virtue.
In the end, it is not the AI‑crafted message that deserves our trust. It is the person who stands behind it. The tool may shape the words and increase their effectiveness, but only the human can supply the integrity.
You cannot trust that almost any creative content isn't AI-generated. There are some fingerprints on stuff that can help identify it as AI-generated, but I expect that to get cleaned up over time.
We could use a name correction here - Artificial Intelligence is inaccurate and, well, insincere - we still have yet to see an example of anything that could even generously be called intelligence, much less a new thought or nonrecombinant contribution. I see sophisticated Boggle and increasingly shrewd clicks and whistles, and that's about it. Normalize the use of a 'Generated Using Artificial Intelligence' tag where AI is exercised, and watch the value proposition of real human writing flourish. We've had super-computers for decades. Even weird language-model supercomputers are just that - not intelligent. I remain unmoved by all of this.
If an organization promotes candor and truth and recognizes subjectivity, then AI generated writing is going to be counterproductive. For organizations that value loyalty and secrecy above just about anything else, then AI generated writing is a useful tool. Alas, most organizations fall or careen between the two extremes depending on leadership or lack thereof.
And then there is the internal corrosion.
Four hundred and one years ago Francis Bacon wrote that “Reading maketh a full man, conference a ready man, and writing an exact man.” Today, scrolling copy alone and having AI generate our writing appears set to prove the effects of the opposite behaviors.
One of the pleasures of writing something, of being clear in words that others will read, is finding out what you are actually thinking. When people use AI for their writing they corrode the quality of their thought.
We’d talk about this more, but too many of us have leaned into AI to form our thoughts effectively.
The Bacon quote is exactly right — and it's exactly the spirit in which I try to use AI: "Conference maketh a ready man." My best AI interactions are conference, not dictation. The AI offers a framing. I push back. It pushes back. Something neither of us started with emerges. The discipline of making my thinking precise enough to evaluate, extend, or reject someone else's contribution is at least as demanding as writing alone. Maybe more, because I'm developing my own position while engaging with another's — even if (or maybe especially if) the other in this case is a non-human intelligence.
The corrosion you're describing is real — but it's a symptom of how most people use AI, not of AI itself. The person who prompts "write me a memo" and sends whatever comes back has skipped the exactness Bacon was talking about. Their thinking never got precise. But that's a choice, not an inevitability.
I've spent months building tools that go the other direction — AI that forces clarity of thought rather than replacing it. Here's one that structures exploratory thinking, requiring you to question your own framing before converging on an answer: https://github.com/mightytech/nla-framework/blob/main/core/skills/think.md. Here's one that forces you to understand and articulate the strongest version of a position you disagree with before arguing against it: https://github.com/mightytech/nla-process-helpers/blob/main/app/steelman.md. And here's one that asks the AI to reflect on what happened in a working session — what worked, what didn't, what the human contributed that the AI couldn't have reached alone: https://github.com/mightytech/nla-framework/blob/main/core/skills/debrief.md.
These aren't AI writing for you. They're AI creating structured spaces where your thinking has to become exact — in Bacon's sense. The tool can corrode thought or discipline it. The difference is whether you're using it to avoid thinking or to think harder than you could alone.
Bacon would recognize the process, I think, even if the medium would surprise him.
"One of the pleasures of writing something, of being clear in words that others will read, is finding out what you are actually thinking. When people use AI for their writing they corrode the quality of their thought." I find the opposite. Note your own words " to form our thoughts effectively." Isn't that the purpose of writing?
I liked this essay, but the author engages in hyperbole when she states that language creates reality. Language emerges from reality; it is our (often feeble and inadequate) response to reality. But the underlying “ground of being” is not a word or even “The Word.” We humans congratulate ourselves on our clever use of words by making such claims, but the entire realm of nonverbal conduct, which often (always?) has more salience gives the lie to this categorical mistake.
I share some of these misgivings. The boss who pastes a ChatGPT congratulations email without reading it is being inauthentic — but not because AI was involved. They'd be equally inauthentic with a Hallmark card or a copied-and-pasted message from last year. The inauthenticity is in the absence of thought, not the presence of a tool.
But the essay defines authenticity as a specific process — the writer alone choosing each word — and treats any other process as inauthentic by definition. A collage artist doesn't create each image fragment from scratch. A photographer doesn't paint each pixel. Nobody calls these media inauthentic because the artist operates at the level of composition and vision rather than individual rendering. AI-assisted writing is a different medium that operates at a different level. The authenticity lives in the thinking, the judgment, the perspective — not in whether one human selected every word.
More importantly, the essay treats all writing as expression. Most writing isn't. It's functional. A single mother disputing a medical bill doesn't need authentic self-expression — she needs clear, legally informed language that protects her rights. Before AI, her options were write it herself in language that might be unclear or legally naive, or pay a lawyer she can't afford. AI gives her a voice she never had in a domain where her natural voice was a disadvantage. That's not inauthenticity. That's access.
A student in Ghana can ask AI to translate an English-language article into Twi and then engage with the ideas in their strongest language. The engagement — their questions, their connections to their own context — is deeply authentic. The translation removed a barrier, not an authenticity.
The hand-wringing about AI and writing comes disproportionately from people whose specific relationship to writing is threatened — journalists and creative writers who experience writing as identity. That's real and worth taking seriously. But most people experience writing as obligation: the email they have to send, the form they have to complete, the application they have to get right in a language that isn't their strongest. For them, AI isn't undoing authenticity. It's removing an obstacle between their intent and their expression. The intent was always there. The craft wasn't. Now it's available.
The real question isn't whether AI-assisted writing can be authentic. It's whether the person using AI bothered to think. That question has the same answer it's always had, with or without AI.
Well said!!!
The entire media economy rests on a single premise: persuading people to part with their money for things of dubious or deceptive value. Advertising.
And the entirety of public discourse has decayed into a parallel enterprise: persuading people to vote for things of similarly dubious or deceptive value. Politics.
Before AI arrived, we had reached a collective conclusion that we could no longer trust anything anyone writes or says.
Enter AI.
This article insists that AI-generated content will destroy trust and authenticity. But how do you destroy what has already been reduced to rubble?
Trust has always been a personal attribute. It belongs to the person, not the tool. If society wants to recover it, we must relearn how to judge character — honesty, clarity, logic, common sense — rather than credentials, celebrity, or the sheer volume of one’s outrage.
And here is the irony: precisely because AI is inhuman, it may help strip away the human pretenses that have corrupted advertising and politics. If it exposes the hollow performances, the manipulations, the failed character behind the message, then authenticity might again become visible — not as branding, but as a personal virtue.
In the end, it is not the AI‑crafted message that deserves our trust. It is the person who stands behind it. The tool may shape the words and increase their effectiveness, but only the human can supply the integrity.
You cannot trust that almost any creative content isn't AI-generated. There are some fingerprints on stuff that can help identify it as AI-generated, but I expect that to get cleaned up over time.
We could use a name correction here - Artificial Intelligence is inaccurate and, well, insincere - we still have yet to see an example of anything that could even generously be called intelligence, much less a new thought or nonrecombinant contribution. I see sophisticated Boggle and increasingly shrewd clicks and whistles, and that's about it. Normalize the use of a 'Generated Using Artificial Intelligence' tag where AI is exercised, and watch the value proposition of real human writing flourish. We've had super-computers for decades. Even weird language-model supercomputers are just that - not intelligent. I remain unmoved by all of this.
If an organization promotes candor and truth and recognizes subjectivity, then AI generated writing is going to be counterproductive. For organizations that value loyalty and secrecy above just about anything else, then AI generated writing is a useful tool. Alas, most organizations fall or careen between the two extremes depending on leadership or lack thereof.