By William Ashley Johnson

This final post in a three-part series explores how artificial intelligence (AI) prompt writing changes when the task is not academic performance or professional communication, but personal expression. In personal writing, the prompt becomes less about producing a polished answer and more about protecting voice, memory, feeling, and intention. The central challenge is not whether AI can help us write something smoother, but whether we can ask for help without letting the tool sand away the human irregularities that made the writing matter in the first place.

Ask AI to write “a heartfelt birthday message for my best friend,” and it will usually hand back something that sounds like it was printed inside a card next to a watercolor balloon. There will be gratitude, laughter, and almost certainly some version of “I am so lucky to have you in my life.” It will be perfectly fine, which is exactly the problem.

Personal writing is rarely built out of perfectly fine.

A real birthday message might mention the time your friend drove forty minutes to bring you fries because you were having a minor emotional collapse over a Tuesday. A real apology might need to admit not only what happened, but what you were too embarrassed to say the first time. A real journal entry may begin with “I don’t know why this bothered me so much,” then spend three messy paragraphs proving that, actually, you do.

That is where personal prompting gets complicated. In academic writing, the prompt helps organize thought. In professional writing, it helps manage audience, purpose, tone, timing, and stakes. In personal writing, the prompt has to do something even stranger: it has to make room for the part of the writing that does not yet know what it means.

That may be why vague personal prompts so often produce language that feels emotionally correct but strangely uninhabited. “Write a touching tribute to my grandmother” may return a smooth meditation on kindness, wisdom, and love. But unless the prompt gives the model the texture of the actual relationship, the smell of her kitchen, the phrase she used every time she got annoyed, the result will hover above the truth instead of landing in it.

Research on AI and creative writing sharpens that concern. Doshi and Hauser (2024) found that access to generative AI ideas can make stories seem more creative, better written, and more enjoyable, especially for writers who begin with lower measures of creativity, but it can also reduce the collective diversity of the stories produced. In other words, AI can help a piece sound better while nudging writers toward similar patterns, ideas, and shapes.

That finding feels especially important for personal writing because sameness is exactly what personal writing cannot afford. A condolence note, a memoir fragment, a wedding toast, a letter to a former self, a journal entry written at midnight because the mind will not settle–these forms depend on particularity. They need the odd detail, the crooked sentence, the emotional contradiction. They need words that would not belong to anyone else.

O’Sullivan’s (2025) stylometric comparison of human and AI-generated creative writing points in a similar direction. The study found that human-authored stories showed broader stylistic variety, while large language model outputs clustered more tightly and showed greater stylistic uniformity. AI writing can be fluent and coherent, but that fluency often arrives with a kind of sameness built into the music of the sentences.

That does not mean AI has no place in personal writing. I think it means the prompt has to become more protective. The writer has to decide what kind of help is actually being requested. “Make this sound better” is risky because “better” often means “smoother,” and smoother is not always more honest. “Help me find the emotional center of this paragraph without changing my voice” is a very different request. So is “Ask me five questions that would help me remember this more clearly before I draft it.” So is “Point out where this sounds generic and where it sounds most like me.”

Those prompts treat AI less like a ghostwriter and more like a patient listener: one who can reflect, question, organize, and nudge without taking over the whole room.

This is why the personal prompt has to include boundaries. Not just “write this for me,” but “do not erase the awkwardness.” Not just “make this emotional,” but “keep the humor dry and understated.” Not just “revise this letter,” but “help me sound honest without sounding cruel.” Personal writing benefits from context, but it also benefits from restraint. The more intimate the writing, the more the writer has to decide what AI is allowed to touch.

I have seen this in my own writing, especially in the early moments when I am not asking for sentences yet. I am asking for pressure. A question. A way back into the material. Sometimes the most useful prompt is not “draft this scene,” but “what are three possible emotional tensions in this memory?” Sometimes, it is even simpler: “Help me make a list of what I remember before I try to shape it.” At that stage, AI is not replacing the personal act of writing. It is helping create a surface for the writer to push against.

That may be the best version of AI-assisted personal writing: not automation, but resistance. The tool gives something back, and the writer gets to say, “No, that is not quite it.” That rejection can be useful. A too-polished sentence can reveal where the actual voice is rougher. A generic tribute can remind the writer of the one detail that refuses to be generic. A bland reflection can make the writer realize the real feeling is not gratitude, but regret, or relief, or love mixed with irritation, which is often where honest writing begins.

Maybe this is where the whole “Power of the Prompt” series has been headed. In academic writing, prompting reminds us that good questions make thinking possible. In professional writing, prompting reminds us that clear requests help words do useful work. In personal writing, prompting reminds us that specificity is not only about efficiency or correctness. Sometimes, specificity is how we protect the human thing at the center of the page.

AI can help with personal writing, but it cannot know what the memory costs. It cannot know why one ordinary detail matters more than the grand summary. It cannot know which sentence sounds like your father, your friend, your younger self, or the person you are still trying to forgive. Those personal echoes, the ones that give the writing its pulse, still have to come from us.

The power of the personal prompt, then, is not that it lets us outsource the self. It is that it asks us to name the self more carefully before we write. What do I mean? What do I remember? What tone is true? What should stay imperfect? What must not be smoothed away?

Those questions are not new. Writers have always asked them. AI simply makes the request visible. And in personal writing, that visibility may be the whole point: before the tool can help us say something, we have to decide what is ours to say.

References

Doshi, A. R., & Hauser, O. P. (2024). Generative AI enhances individual creativity but reduces the collective diversity of novel content. Science Advances, 10(28), eadn5290. https://doi.org/10.1126/sciadv.adn5290

O’Sullivan, J. (2025). Stylometric comparisons of human versus AI-generated creative writing. Humanities and Social Sciences Communications, 12, Article 1708. https://doi.org/10.1057/s41599-025-05986-3 

---

About the Author

William Ashley Johnson teaches English composition and rhetoric at Purdue Global, where his work explores the intersections of user experience (UX), writing pedagogy, and generative AI in online learning environments.