Audio Overview
ENHow to Avoid AI-isms in Your Writing
AI is here to stay, and writing is one of its most useful applications. There is little point in refusing a useful tool simply to prove that we can work without it. We use calculators even though we understand arithmetic, and spreadsheets even though we know how the calculations behind them work. Using AI to help us write should be no different. The tool can help with the work without replacing the thinking behind it.
But this piece is not about using a single prompt to generate something we have barely thought about ourselves. It is for knowledge workers who have already done the thinking. These are people writing an argument, an analysis, a proposal, an article, or anything else in which their judgement and ideas are the real source of value.
AI can be extremely useful in turning those thoughts into a well-structured piece of writing. The problem is that, while doing so, it often introduces its own recognisable habits, including certain sentence constructions, words, punctuation, rhythms and rhetorical devices that increasingly signal “AI wrote this.”
That matters because the reader cannot see how much thinking went into the work before AI became involved. A thoughtful piece can contain original ideas, careful reasoning and genuine judgement, yet a few recognisable AI-isms can make it look like just another piece of AI slop. The risk is that the writing style ends up obscuring the very thing that makes the work valuable, the author's thinking behind it.
The answer is not to stop using AI. It is to learn to recognise the fingerprints it tends to leave behind.
Here are some of the easiest ones to spot.
| Category | AI-ism | What to watch for | Example |
|---|---|---|---|
| Sentence structure | Negative parallelism | Repeatedly defining something by saying what it isn't before saying what it is. | “It’s not about working harder. It’s about working smarter.” |
| Sentence structure | Rhetorical question + immediate answer | Asking a question only to answer it in the very next sentence. | “So what does this mean for businesses? It means they need to adapt.” |
| Reasoning | False causal connection | Presenting one statement as the cause or explanation of another when the connection has not actually been established. | “The natural dam broke and released the river. As a result, the gorge became one of the deepest in the world.” |
| Emphasis | Empty emphasis | Announcing that something is important instead of showing why it is important. | “It’s important to remember that…” / “It’s worth noting that…” |
| Opening | Stock scene-setting | Beginning with a broad, generic statement that could introduce almost any article on the subject. | “In today’s rapidly evolving digital landscape…” |
| Vocabulary | AI-favoured vocabulary | Noticeably polished words that recur frequently in AI prose. | “delve,” “underscore,” “pivotal,” “multifaceted,” “tapestry,” “realm” |
| Vocabulary | Buzzword inflation | Using grand business language where an ordinary verb would say the same thing. | “Harness the power of AI to unlock your organisation’s full potential.” |
| Vocabulary | Synonym stacking | Piling up words with essentially the same meaning to make a statement sound stronger. | “This represents a major, significant and transformative shift.” |
| Punctuation | Em-dash overuse | Repeatedly inserting thoughts between em dashes instead of writing straightforward sentences. | “AI — when used thoughtfully — can help knowledge workers — particularly those under time pressure — communicate more effectively.” |
| Rhythm | Tricolon / rule of three | Repeatedly packaging ideas into neat groups of three for rhetorical effect. | “Work faster, think smarter, achieve more.” |
| Rhythm | Excessive parallelism | Repeating the same grammatical construction to create an unnaturally polished rhythm. | “Designed to simplify. Built to scale. Engineered to transform.” |
| Conclusion | Formulaic summary | Announcing the conclusion with a stock phrase rather than simply making the concluding point. | “In conclusion…” / “In short…” / “At the end of the day…” |
| Conclusion | Generic inspirational ending | Finishing with a polished, sweeping statement that sounds meaningful but adds little. | “The future isn’t something we wait for. It’s something we create.” |
A prompt you can use
Knowing what to look for makes these patterns much easier to catch when reviewing a draft. But it is better if fewer of them appear in the first place. The following prompt can be added to your instructions when you give AI your ideas, arguments and source material. It tells the AI to help turn your thinking into clear writing without introducing the habits in the checklist above.
Quick explainer
Paste this prompt into the context window
Using the information, ideas and arguments I provide, write the piece I ask for while preserving my reasoning, meaning and point of view. Help with structure, clarity and flow, but do not replace my thinking with generic AI-generated ideas or embellishment. Keep the writing direct, natural and specific. Avoid recognisable AI writing habits, including: Negative parallelism such as “It’s not about X. It’s about Y.” Rhetorical questions that are immediately answered. False causal connections. Do not imply that one fact causes or explains another unless the information I provided supports that connection. Empty emphasis such as “It’s important to remember” or “It’s worth noting.” Stock scene-setting such as “In today’s rapidly evolving landscape.” AI-favoured words such as “delve,” “underscore,” “pivotal,” “multifaceted,” “tapestry” and “realm” when simpler language works. Inflated phrases such as “harness the power of” and “unlock the potential of.” Stacking synonyms or similar ideas merely for emphasis. Unnecessary or excessive em dashes. Repeated groups of three or slogan-like constructions such as “faster, smarter, better.” Excessive parallel sentence structures such as “Designed to simplify. Built to scale. Engineered to transform.” Formulaic conclusions such as “In conclusion,” “In short” and “At the end of the day.” Generic inspirational endings that sound polished but add no substantive meaning. Do not invent facts, reasoning, examples or conclusions simply to make the writing sound more complete. Where my reasoning does not support a connection or conclusion, do not manufacture one. The finished writing should sound like a thoughtful person expressing ideas they have genuinely worked through, with AI helping them communicate those ideas clearly.
Using AI to help you write does not make the thinking any less yours, provided the thinking really did come from you. The problem begins when AI's habits become more visible than the ideas you brought to the page.
Use AI for what it is good at. Let it help you organise, clarify and express your thoughts. Then remove the fingerprints it leaves behind, so what the reader notices is your thinking, not the tool that helped you write it.
How to Avoid AI-isms in Your Writing
AI is here to stay, and writing is one of its most useful applications. There is little point in refusing a useful tool simply to prove that we can work without it. We use calculators even though we understand arithmetic, and spreadsheets even though we know how the calculations behind them work. Using AI to help us write should be no different. The tool can help with the work without replacing the thinking behind it.
But this piece is not about using a single prompt to generate something we have barely thought about ourselves. It is for knowledge workers who have already done the thinking. These are people writing an argument, an analysis, a proposal, an article, or anything else in which their judgement and ideas are the real source of value.
AI can be extremely useful in turning those thoughts into a well-structured piece of writing. The problem is that, while doing so, it often introduces its own recognisable habits, including certain sentence constructions, words, punctuation, rhythms and rhetorical devices that increasingly signal “AI wrote this.”
That matters because the reader cannot see how much thinking went into the work before AI became involved. A thoughtful piece can contain original ideas, careful reasoning and genuine judgement, yet a few recognisable AI-isms can make it look like just another piece of AI slop. The risk is that the writing style ends up obscuring the very thing that makes the work valuable, the author's thinking behind it.
The answer is not to stop using AI. It is to learn to recognise the fingerprints it tends to leave behind.
Here are some of the easiest ones to spot.
| Category | AI-ism | What to watch for | Example |
|---|---|---|---|
| Sentence structure | Negative parallelism | Repeatedly defining something by saying what it isn't before saying what it is. | “It’s not about working harder. It’s about working smarter.” |
| Sentence structure | Rhetorical question + immediate answer | Asking a question only to answer it in the very next sentence. | “So what does this mean for businesses? It means they need to adapt.” |
| Reasoning | False causal connection | Presenting one statement as the cause or explanation of another when the connection has not actually been established. | “The natural dam broke and released the river. As a result, the gorge became one of the deepest in the world.” |
| Emphasis | Empty emphasis | Announcing that something is important instead of showing why it is important. | “It’s important to remember that…” / “It’s worth noting that…” |
| Opening | Stock scene-setting | Beginning with a broad, generic statement that could introduce almost any article on the subject. | “In today’s rapidly evolving digital landscape…” |
| Vocabulary | AI-favoured vocabulary | Noticeably polished words that recur frequently in AI prose. | “delve,” “underscore,” “pivotal,” “multifaceted,” “tapestry,” “realm” |
| Vocabulary | Buzzword inflation | Using grand business language where an ordinary verb would say the same thing. | “Harness the power of AI to unlock your organisation’s full potential.” |
| Vocabulary | Synonym stacking | Piling up words with essentially the same meaning to make a statement sound stronger. | “This represents a major, significant and transformative shift.” |
| Punctuation | Em-dash overuse | Repeatedly inserting thoughts between em dashes instead of writing straightforward sentences. | “AI — when used thoughtfully — can help knowledge workers — particularly those under time pressure — communicate more effectively.” |
| Rhythm | Tricolon / rule of three | Repeatedly packaging ideas into neat groups of three for rhetorical effect. | “Work faster, think smarter, achieve more.” |
| Rhythm | Excessive parallelism | Repeating the same grammatical construction to create an unnaturally polished rhythm. | “Designed to simplify. Built to scale. Engineered to transform.” |
| Conclusion | Formulaic summary | Announcing the conclusion with a stock phrase rather than simply making the concluding point. | “In conclusion…” / “In short…” / “At the end of the day…” |
| Conclusion | Generic inspirational ending | Finishing with a polished, sweeping statement that sounds meaningful but adds little. | “The future isn’t something we wait for. It’s something we create.” |
A prompt you can use
Knowing what to look for makes these patterns much easier to catch when reviewing a draft. But it is better if fewer of them appear in the first place. The following prompt can be added to your instructions when you give AI your ideas, arguments and source material. It tells the AI to help turn your thinking into clear writing without introducing the habits in the checklist above.
Quick explainer
Paste this prompt into the context window
Using the information, ideas and arguments I provide, write the piece I ask for while preserving my reasoning, meaning and point of view. Help with structure, clarity and flow, but do not replace my thinking with generic AI-generated ideas or embellishment. Keep the writing direct, natural and specific. Avoid recognisable AI writing habits, including: Negative parallelism such as “It’s not about X. It’s about Y.” Rhetorical questions that are immediately answered. False causal connections. Do not imply that one fact causes or explains another unless the information I provided supports that connection. Empty emphasis such as “It’s important to remember” or “It’s worth noting.” Stock scene-setting such as “In today’s rapidly evolving landscape.” AI-favoured words such as “delve,” “underscore,” “pivotal,” “multifaceted,” “tapestry” and “realm” when simpler language works. Inflated phrases such as “harness the power of” and “unlock the potential of.” Stacking synonyms or similar ideas merely for emphasis. Unnecessary or excessive em dashes. Repeated groups of three or slogan-like constructions such as “faster, smarter, better.” Excessive parallel sentence structures such as “Designed to simplify. Built to scale. Engineered to transform.” Formulaic conclusions such as “In conclusion,” “In short” and “At the end of the day.” Generic inspirational endings that sound polished but add no substantive meaning. Do not invent facts, reasoning, examples or conclusions simply to make the writing sound more complete. Where my reasoning does not support a connection or conclusion, do not manufacture one. The finished writing should sound like a thoughtful person expressing ideas they have genuinely worked through, with AI helping them communicate those ideas clearly.
Using AI to help you write does not make the thinking any less yours, provided the thinking really did come from you. The problem begins when AI's habits become more visible than the ideas you brought to the page.
Use AI for what it is good at. Let it help you organise, clarify and express your thoughts. Then remove the fingerprints it leaves behind, so what the reader notices is your thinking, not the tool that helped you write it.
Continue reading