By skyforbes 
I ran into a pretty common but weird dilemma that I’m sure some of you university folks will recognize: I’ve got a professor who’s secretly (or maybe not-so-secretly) not into anything that smells like it came from ChatGPT. He’s a good prof, don’t get me wrong, but every time he catches the faint whiff of “AI-essay” vibes, he either changes the assignment last minute or gives the work an unusually thorough scan. So I found myself thinking: how do you still get the benefits of AI assistance (drafts, ideas, structure) while making sure the final submission reads like it was done by you, not a machine?
Here’s how I approached it, with real scenarios, lessons learned, and one tool I found surprisingly helpful (spoiler: it's not just more AI).
🎓 Real scenarios from my college life
- Midnight research panic End-of-term week, I have a 2,000-word essay due in 18 hours. ChatGPT gives me an outline and a draft in 10 minutes. But when I handed it in as is, I got feedback like: “Voice too generic; doesn’t feel like you.”
- Fix: I reopened the draft, added personal anecdote about when I tried the experiment in class, changed phrasing to how I talk, not how “the researcher” talks.
-
Result: Prof commented “Better tone; good personal insight.” ✅ -
**Group project with uneven members** My group used ChatGPT to generate the first version of our ‘industry trends’ slide deck. One member basically copied it verbatim. Prof flagged the presentation as “looks like from a generic business blog” and asked us to re-do. -
Fix: We took the draft, inserted our own experiences (internship, part-time job), changed bullet phrasing to “We noticed…” instead of “It is observed that…”, and added a funny slide of our team photo in hoodies. -
Result: Audience laughed, prof approved. Our voice passed. -
**Discussion board thread where the prof reads everything** Every week the professor reads all replies posted to the discussion board. One time I used ChatGPT to craft a reply, but it sounded too perfect (“The salient point illustrated…”) and I got a “See me after class” message. -
Fix: Next time I used the AI draft for structure, but rewrote into conversational tone: “Cool question—here’s my take…” Then mentioned something from class earlier that week to link to me. -
Result: No red flags. Prof even quoted my post in the lecture.
How to Humanize the AI Text
Here’s a checklist I now follow every time I use AI-assisted writing:
-
**Use it for ideas, not final phrasing.** Let ChatGPT give you the outline, bullet points, maybe draft paragraphs—but don’t hand it in verbatim. -
**Insert your voice.** Think about how you talk in class, in messages, or to friends. Do you say “for sure” or “indeed”? Use that tone. Add small details only you know. -
**Add personal context or story.** Even a 1-sentence example from your own experience makes the text feel unique. Example: “When I ran the lab test on Wednesday…” or “Last semester I thought…” -
**Check for artificial phrasing.** ChatGPT sometimes uses sudden formal transitions or odd words. Replace anything that feels “out of you.” E.g., change “Henceforth” to “So…”. -
**Read aloud & vary sentence length.** Real human writing has short, medium, long sentences... AI drafts often go steady-bath-ready. Read aloud: if it feels robotic, change it. -
**Use a tool like Grubby AI for tone adjustments.** I discovered Grubby AI (yes, first time hearing of it too) and found it useful. Here’s how: -
I paste the AI text into Grubby, choose “Casual / University tone”, ask it to “make it sound like a student.” -
Then I still edit it. Grubby helps bridge the gap between “too generic AI” and “definitely me.” -
It reduces phrases like “in the aforementioned manner” and replaces with “So yeah, I think…”. -
Final check: Run it by a peer or your own voice. If you cringe reading a sentence like you’re not the one saying it — change it. If your friend says “sounds like you” vs “looks like it was done by someone else” — that’s a good sign.
🎥 For more tips:
Here’s a video I found helpful (makes the point clearer than I can here): https://www.youtube.com/watch?v=ltqHxgJcuDQ&t=1s
✅ Why it matters
-
When your professor hates AI-only submissions, you reduce risk of it being flagged. -
You keep the productivity boost of AI for ideas, structure, first draft. -
Your writing still sounds like you, so you’re more likely to get good engagement, better marks, and fewer pitfalls.