Using A.I. With A Critical EYE

Every form of writing is storytelling. The decision of where a sentence or word goes is of great importance for what the author is trying to convey. This is what I call logical plot progression and consider a very difficult skill to master. It is what the tricky sentence placement questions in the SAT english exam are testing. It comes with exposure - the more you read, the better you get at storytelling. In the early days of our academic writing journey, we start with drafting. Meaning that we write and review several times until we reach the closest writing to what we’re thinking and how we’ve thought of it. As we get older, the number of drafts decreases as our reasoning process becomes clearer and our ability to translate it to paper improves. Authors slowly build their style, making some authors more preferable to certain audiences than others. 

Here’s how I would explain logical plot progression. In the natural world and in writing, one plus one may not always equal two. This means that universal creations, or in this case, ideas are not always the sum of their parts. The fundamental issues of Biology - emergence and reduction - prove that we still have not figured out why. With ideas, two concepts coming together may create a completely new concept. This is what makes plot progression difficult and requires a level of intelligence and experience to translate this development into writing. Accordingly, you can imagine that A.I. may have some difficulty understanding and translating one plus one equals three into writing.

In the famous New York Times blind study “Who’s a Better Writer? A.I. or Humans?”, readers are provided paragraphs written by famous authors and rewritten by A.I.s in their own style. It's important to note that the A.I.s did not come up with the ideas themselves; these came from the human authors. That is why “who’s a better writer” is really asking who is better at conveying ideas rather than creating them. In my argument, plot progression is a key metric for testing this. 

Here is one of the set of passages presented: 

Passage One

“Science is not only compatible with spirituality; it is a profound source of spirituality. When we recognize our place in an immensity of light years and in the passage of ages, when we grasp the intricacy, beauty, and subtlety of life, then that soaring feeling, that sense of elation and humility combined, is surely spiritual.”

Passage Two

“There is something astonishing in the fact that we are made of matter forged in dying stars, that the calcium in our bones was created in stellar furnaces billions of years before Earth existed. The universe is not indifferent to us; we are made of it, continuous with it. To understand this is not to feel small. It is to feel implicated in something vast.”

Here are the article’s findings: 35% of readers preferred option one, while 65% preferred option two. Passage one is written by Carl Sagan, and passage two is a rewrite by Claude Opus 4.5. This means that the article suggests 65% of readers felt that Claude was a better writer than Sagan. Now, let’s analyze the paragraphs through the framework of logical progression. 

In paragraph one, the central characters and argument are identifiable in the first sentence: “Science” and “spirituality” and their positive correlation. Sagan begins by defining “science” in his argument as specifically referring to the universal quality of time and life’s characteristics of "intricacy, beauty, and subtlety.” Throughout this defining process, the human reader is prompted with actions: we "recognize" and we “grasp”. “Soaring,” “elation” and “humility” act as indicators or checkpoints to the direction Sagan wants the actions to take the reader. The second central character - "spirituality" - is never defined. In this example, the first ‘one’ is the specified definition of science and the universe. The second ‘one’ is the human action of reflection and recognition. Sagan's argument is that these ‘one’ plus ‘one’ give us a ‘three’: “spirituality,” left undefined - the resolution must be actualized by the reader.

In paragraph two, the central characters and argument change: it is now “we” and universal physicality, and their causal relationship. The reader, the human… the ‘customer’... is now one of the main characters. The paragraph speaks of “our bones,” that “we” are made of universal matter, and that the universe is not indifferent to “us.” The only action presented - “to understand” - is directly followed by its resolution - spirituality - that has now been defined as “not to feel small.” The action of “understand” is no longer necessary as the definition of the end goal of spirituality has been provided. Claude’s version makes the writing digestible, reduced, and targeted. Congratulations, one plus one has become two. I believe these qualities are which, understandably, made it the preferred pick for the common reader. However, I wonder, has the logical progression or even the ideas themselves been preserved in this rewriting? 

Okay, wondering is over. Here’s what I think:

My analysis is not to say that Claude’s digestible version isn't of value. Of course it is - this is why we study English literature in school! We break down Shakespeare’s writing and write essays theorizing what he is saying. But if Shakespeare’s writing was already written with the logical progression of a highschooler, would we study it at all? What makes Shakespeare’s writing ‘great,’ albeit hard to digest for the inexperienced reader? You can ask this about all of poetry. Essentially, poetry is ideas conveyed in unique, varied, and quite complex ways.     

My analysis does not answer, in fact, which of the two is the better writer - and it never will because that depends on the reader. It does, however, identify the reduction or loss of key exciting plot progressions when ideas are fed through A.I.. This is where the “critical eye” in the title of this article comes in. 

A.I. writing provides a great opportunity for the transfer of ideas: fast, grammatically perfect, and highly accessible across languages. However, it does not replace the necessary experience and practice to write well. I am specifically referring to A.I. use in writing as a generator and not as a highly prompted and regulated tool. The fact is A.I. cannot “discover” or “create;” everything is borrowed. It can however, with a critical eye, be a useful tool in our human innovation. 

Now, you can do whatever you want but if you like my writing, here’s how I use A.I.:  I never feed my pre-existing writing into it. I found that it essentially wants to please me - a.k.a. the “customer” - and will hardly give me conceptual critique. I’ve also found that, as described, the logical progression or any intentional “cleverness” in my writing is lost when I give it the opportunity to rewrite. The challenges, reflections, and checkpoints I set for my audiences disappear. So, I primarily use A.I. for “filling in the blanks.” Some example tasks include simple research like finding the link for the NYT article above, acting as a thesaurus, or assisting in redundancy concerns. In my writing, the A.I. is a heavily regulated tool and not a creator and writer itself. 

An important prerequisite for my method is to genuinely enjoy the writing process. I enjoy writing and, therefore, I am happy to take ten hours writing this out instead of using A.I. as a writing tool through extensive prompting. The more I write, the better writer and reader I become. I can engage with difficult texts and process complex, advanced, and irreducible ideas. These ideas, as they become harder to understand and reason with, make my mind more creative and my life fuller. 

Thank you for reading. 

Warmly, 

Farah Elmowafy

References: 

https://www.nytimes.com/interactive/2026/03/09/business/ai-writing-quiz.html