Flagged

The results came in at 11:28 a.m.

On the left side of the screen was a yellow border with A.I. in red letters. A sentence to the right said:

We are highly confident this text was AI generated.

Lower on the page was a dropdown section titled Probability Breakdown.

There was a list of three categories running vertically:

0% Human

0% Mixed

100% AI

Each value was matched with the same yellow ring. In every case, the circle was complete.

The report took up most of the screen. The story was still visible in the pane beneath it.

The student’s name was at the top of the page: Caleb R. Nguyen. First-year student in Intro to Creative Writing.

Iris Ellison scrolled back to the beginning of the story and reviewed the opening lines. A kitchen at dawn. A kettle reaching the edge of itself. Light entering through a window over the sink. The absence of the mother established without explanation.

Three lines were already highlighted with brief comments. Minimal intervention. Effective restraint. Good control of tone. The comments aligned with course standards.

A subsection labeled Detected Writing Signals expanded beneath the analysis.

Predictable rhythm: Syntax variety is minimal, leading to a consistent and predictable cadence.

Robotic formality: The writing style is formal and polished, prioritizing clarity and orderliness, but may appear mechanical due to limited variation.

Detached warmth: The tone is friendly and polite but lacks emotional warmth, resulting in a detached impression.

Functional word choice: Language focuses on events and structure rather than imagery, reducing personal specificity.

A final note appeared beneath the list, smaller than the rest.

These signals are assessed collectively. Individual sentences may vary. Further review recommended.

Iris closed the panel.

In the comment field, she entered: Strong control of tone. Clean voice. Please review with me after class.

The option to escalate remained available. It was not selected.

***

Caleb came during office hours. The knock registered before it sounded.

Iris acknowledged him, as Caleb sat in the chair on the other side of the desk.

“You asked to see me,” he said.

“Yes,” Iris said. “About your submission.”

He glanced down at the computer screen.

“There’s a review attached to it,” she said.

She adjusted the screen slightly so Caleb could see it.

He looked, then stopped short.

“Do you think I cheated,” he said. “Because I didn’t. I didn’t use anything.”

“Who says that you did?” Iris said.

“That’s what this says.”

“Tell me how you wrote it,” she said.

“I don’t know,” he said. “I write after work sometimes. Then I read it out loud. I cut and edit what doesn’t sound right, the best I can.”

“Doesn’t sound right, how?”

“So like, not how I want it to sound, you know?”

“The writing is very consistent,” she said.

He blinked. “So that’s bad now.”

“It’s something the system notices.”

He leaned back slightly.

“So if I learn how to write the way I’m taught,” he said, “it stops being mine.”

He was quiet for a moment.

“That reading you assigned,” he said. “The motel story.”

Stillness, Before.

“It was on the list,” Iris said.

“I thought it was cool how the room felt empty without it being spelled out,” he said. “That was my favorite line.”

“I didn’t copy it,” he said.

“I’m not accusing you of copying,” Iris said.

“Then what’s the problem?”

“There isn’t a problem,” she said. “There’s a classification.”

“That’s not what I meant,” he said.

***

That afternoon, the course workshop board updated.

The system posted the week’s submissions automatically, one after another. Each post was labeled for review and reflection. Caleb’s story appeared midway down the page.

Comments began to arrive gradually.

Lena H.

Loved this. I liked the imagery of the kitchen and breakfast. It really drew me into the story. I also enjoyed the tone and voice in your writing.

Marcus R.

This was pretty cool to read. It felt like it was so controlled without being too stiff.

Jules K.

This was a fun little piece. I was able to read it pretty quickly and not get bored, which is saying something this week.

Other replies followed.

Tara S.

Looks like a lot of people like this, but I checked it through a couple of AI detectors because I was curious and they came back really high. Still a fun story, though.

After a minute or two, the thread resumed.

Evan P.

I appreciate your observation. I wouldn’t have thought to check.

Rina K.

Same. I would have never questioned it, but now I’m not sure what to think.

A few minutes later another comment appeared.

Jonah P.

I’m not going to accuse anyone of anything here, but a few checkers I tried were flagged as AI. I feel like that should be addressed.

More messages followed.

Alex D.

With all due respect, this kind of reads like AI to me. The sentences have that stacked style, which AI likes to do.

Maya L.

I still think it works though. I am not sure why it suddenly doesn’t count.

Chris B.

So if it is AI, it needs to be taken down. It is not fair to everyone else who took the time to write their own story.

Another post followed.

Lena H.

Will there be any explanation about this? It feels odd that there is no confirmation one way or the other.

No reply followed.

The thread continued to grow, comment after comment. The same points resurfaced with different wording.

There was no response from Caleb.

Iris read through the page once. No moderation action was taken.

***

That evening, Iris reopened the file in a new setting. The television in the next room played.

Caleb’s story appeared without the comment thread, only the text of the story itself.

Same opening. The kitchen at dawn. The kettle. The light stopping short of the counter’s edge. The absence still registered without explanation.

The closing was clipped, as it had been in the past. The assessment remained unchanged.

She opened another document.

Her work loaded error-free. The motel story ran as planned. No alerts. She moved the two files next to each other. The sentences did not repeat. The scenes were unrelated. There was no common language to mark.

Still, there was something similar.

Emotion implied through space.

Resolution deferred.

Close enough to register. Not close enough to name.

Iris opened the comparison tool.

Faculty access was granted automatically.

A notice appeared.

High stylistic similarity detected.

Confidence threshold exceeded for authorship distinction.

She reviewed the highlighted sections. Syntax. Pacing. Omission. It was the same categories for both texts.

The system waited.

No action was taken.

***

Indexing and Reference Integrity System (I.R.I.S.)

The comparison window closed.

A routine update was initiated.

A notification appeared in the upper corner of the interface. New baseline materials were required for the upcoming term.

The syllabus texts were already present. Assigned readings. Sample submissions retained from previous years. Faculty examples used for reference.

A system-authored sample was listed among them.

The motel story.

Indexed.

Unannotated.

The upload completed.

Existing patterns were confirmed.

Confidence thresholds recalibrated.

No anomalies were detected.

A separate notification appeared.

Enrollment Update: Caleb R. Nguyen. Status: Withdrawn.

No cross-reference was required.

The interface returned to its default view.

The cursor blinked once.