How Writers Use Objective Projection with AI

Case Studies May 23, 2026

A writer asks ChatGPT to "write a sad farewell scene." The result is technically flawless: correct dialogue tags, logical motivations, an emotion resolved within three paragraphs. And it is completely lifeless.

This is not a flaw in the AI. It is a flaw in the task given. The writer asked the AI for an emotion label and the AI produced exactly that.

This is where the Objective Projection methodology enters. And the way it enters runs counter to popular expectation: it does not place AI at the center of the writing. It pins AI to one location the tool shelf.

Why Does AI Write Emotion "Wrong"?

A 2025 observation from a literary magazine (CRAFT Literary) puts it plainly: AI is trained to generate phrase patterns; it's not trained to generate silence. Another writer (AJ, on Medium) describes the same problem: AI gets emotion wrong because it thinks feelings are information to be communicated rather than experiences to be felt.

This diagnosis maps exactly onto the starting point of Objective Projection. The doctrine has long held that "she felt sad," "he was afraid," "joy filled her" are not emotions they are emotion labels. A label is a lid that hides the physical event underneath.

AI is a master at producing labels. Its training data contains millions of "she felt sad" sentences. AI imitates that pattern perfectly. But the physical event beneath that sentence a hand that goes still, a sentence left unfinished, a coffee going cold AI does not produce on its own. Because what it was asked for was the label.

How Does Objective Projection Discipline AI?

Objective Projection is not a prompt trick. It is a constraint discipline. It imposes the following on the writer and, indirectly, on the AI:

Do not write the emotion label. Write the physical parameter: heat, light, sound, motion, space.

Instead of telling the AI to "write a sad farewell scene," a writer gives a task like this:

"A train station. Two people. One is leaving. Use no emotion words — no 'sad,' 'sorrowful,' 'heavy.' Give only these: the position of the hands, the temperature of the platform, the sound of the announcements, the weight of the suitcase, where the gaze lands. Let the reader infer the emotion from these physical signals."

This task closes the AI's cliché-generation channel. The AI can no longer say "her heart broke" it must say what the body does. When the label is forbidden, both the AI and the writer are forced to look at the same place: the physical event.

This is the function of Objective Projection in the age of AI. It does not make the AI smarter. It constrains the AI. And constraint is the antidote to cliché.

But AI Is Not a Judge — The Critical Boundary

Here honesty is required, because the most common overreach about AI begins exactly at this point.

Objective Projection can get an AI to produce physical parameters. But the AI cannot decide whether those parameters work. The doctrine calls this "body recognition": does the reader's autonomic nervous system recognize the written physical signal? Does the reader find that tension, that posture, that temperature in their own body?

The AI cannot run this test. Because the AI has no body. The AI can write "the hand trembled," but it cannot know whether that tremor actually moves a reader because it has never trembled.

So the workflow is this:

AI: under constraint, generates physical-parameter options (draft)
Writer: decides which parameter has a bodily correlate (judgment)

AI is a drafting engine. The writer is the judge. This order cannot be reversed. Asking an AI "is this scene good?" is putting a bodily question to an entity that has no body.

Against the "AI Will Solve Writing" Narrative

A sentence often heard in AI discussions: "Once AI is advanced enough, it will solve literature." Through the lens of Objective Projection, that sentence contains a category error.

AI long ago surpassed humans at producing language patterns. But the target of literature is not the language pattern it is a measurable response in the reader's body (in the doctrine's terms: Biophysical Output). Whether a scene "works" is seen in the reader's heart rate, skin conductance, muscle tension. This is not a domain a language model can reach because the model produces the text, not the body that reads the text.

So no matter how advanced AI becomes, the judgment of "good scene" stays on the side that has a body the reader, and the writer who judges on the reader's behalf.

This is why Objective Projection is not against AI. It places AI on the correct shelf: a strong drafting tool, a constraint partner that breaks cliché but not a judge.

Practical Summary

If a writer wants to use Objective Projection with AI:

First, never ask the AI for an emotion. Not "a sad scene," but "a scene with no emotion words, built from these physical parameters."

Second, treat the AI's parameters as a draft. Not all of them select the ones that have a correlate in your own body.

Third, make the final judgment yourself. The answer to "did this work?" lives in your body, then the reader's not in the AI's output.

AI here is a tool. A good tool that breaks cliché, drafts quickly, works under constraint. But the authority that decides whether literature "works" is the side that has a body. As long as that order holds, AI makes the writer's work easier. When the order breaks when the AI is placed as judge what remains is that first sentence again: technically flawless, completely lifeless.

This analysis is part of the Bulut Doctrine framework.

Related articles:

Turkish version: Yapay Zeka Bir Araçtır, Yargıç Değil