AI detectors - warning

[Anonymous]

Anonymous wrote:My kid used Claude (paid) to cut 42 words from the common app essay. Also used Grammarly to correct grammar (and add commas, semi-colons, broke long sentences up, new dashes, etc).

Beware of Grammarly: The Grammarly changes trigger AI detectors on Pangram, gptzero, Originality, and Winston. The Claude changes did not trigger AI detectors anywhere.

There is increasing chatter on Reddit that AOs have integrated these 3 or 4 third-party AI detectors into "Slate" and other parts of their admissions workflows to screen application materials. Do some research before using Grammarly. It looks like an essay will be auto-highlighted for suspected AI and the entire app thrown out (at some schools).

https://www.enrollify.org/blog/college-ai-detectors
https://www.yomu.ai/resources/do-college-admission-officers-use-ai-to-detect-ai-written-essays
https://www.carnegiehighered.com/blog/slate-ai-features/ (this is crazy!!)
https://knowledge.technolutions.net/docs/summit-2025-executive-summary (click on the dashboard portal!)
https://www.youtube.com/watch?v=qR9b_NGFnt4

Important question: is the chatter FROM the AOs or is it random people surmising that they are doing this?

[Anonymous]

Anonymous wrote:I really don’t understand why any kid needs to use AI at all to write this essay. Not Grammarly, not Claude, not anything. Just write the essay and ask another human to proofread or give feedback.

They never learned how to write properly, probably public school kids.

[Anonymous]

Anonymous wrote:Our kids AO reached out and asked for proof they wrote the essay. My kid sent handwritten rough drafts as proof.

I call BS!!

[Anonymous]

Anonymous wrote:

Anonymous wrote:My kid used Claude (paid) to cut 42 words from the common app essay. Also used Grammarly to correct grammar (and add commas, semi-colons, broke long sentences up, new dashes, etc).

Beware of Grammarly: The Grammarly changes trigger AI detectors on Pangram, gptzero, Originality, and Winston. The Claude changes did not trigger AI detectors anywhere.

There is increasing chatter on Reddit that AOs have integrated these 3 or 4 third-party AI detectors into "Slate" and other parts of their admissions workflows to screen application materials. Do some research before using Grammarly. It looks like an essay will be auto-highlighted for suspected AI and the entire app thrown out (at some schools).

https://www.enrollify.org/blog/college-ai-detectors
https://www.yomu.ai/resources/do-college-admission-officers-use-ai-to-detect-ai-written-essays
https://www.carnegiehighered.com/blog/slate-ai-features/ (this is crazy!!)
https://knowledge.technolutions.net/docs/summit-2025-executive-summary (click on the dashboard portal!)
https://www.youtube.com/watch?v=qR9b_NGFnt4

Important question: is the chatter FROM the AOs or is it random people surmising that they are doing this?

Some of those links show how HYPSM handle (they are the most critical) of any AI.

[Anonymous]

Anonymous wrote:

Anonymous wrote:A lot of intel in these Slate links. Went down the google rabbit hole (if you are interested there's so much more there). Here is what it looks like:

1. Slate can tell if you are "gaming" your major from when you first submitted an online inquiry - bc your two entries are side by side.
https://technolutions.com/admissions?itemId=rlmidt9qexf1yde9xo63qfwrhbby9n


2. What does "gifts" mean?
https://technolutions.com/admissions?itemId=rlmidt9qexf1yde9xo63qfwrhbby9n

3. They are tracking majors by round - even for schools that don't admit by major????
https://technolutions.com/admissions?itemId=y9pht9rumtfmy43uc36m5eep1zsojj


https://technolutions.com/admissions

Not all colleges uses all these "features." In many cases, they have to pay extra to access everything, so don't assume all admissions offices use these fields/features.

I'd assume any T30 private is using these features AND any private that tracks demonstrated interest....bc its a super easy cheat sheet.

[Anonymous]

Anonymous wrote:

Anonymous wrote:I really don’t understand why any kid needs to use AI at all to write this essay. Not Grammarly, not Claude, not anything. Just write the essay and ask another human to proofread or give feedback.

They never learned how to write properly, probably public school kids.

My public school kid wrote a great essay without AI. And your sentence is poorly written.

[Anonymous]

What if they use Grammarly but not the AI feature? Ds uses it but thinks the AI part sucks so he doesn't use it.

[Anonymous]

Anonymous wrote:

Anonymous wrote:Our kids AO reached out and asked for proof they wrote the essay. My kid sent handwritten rough drafts as proof.

I call BS!!

Ha ha - this.

[Anonymous]

Anonymous wrote:Our kids AO reached out and asked for proof they wrote the essay. My kid sent handwritten rough drafts as proof.

This is a lie. THERE IS NOT ONE AO in American that would do this.

[Anonymous]

Anonymous wrote:

Anonymous wrote:Our kids AO reached out and asked for proof they wrote the essay. My kid sent handwritten rough drafts as proof.

this is not believable.

+1

[Anonymous]

Any kid worth $.02 knows to run their essay through the AI detectors and just change enough words until they read 100% human.

I mean…it’s a great first test to weed out the idiots who just submit an AI-generated essay.

[Anonymous]

Anonymous wrote:What if they use Grammarly but not the AI feature? Ds uses it but thinks the AI part sucks so he doesn't use it.

[Anonymous]

I work in this space and we've had 100% human written content flagged as AI and AI content not flagged. None of it is accurate

[Anonymous]

Anonymous wrote:I work in this space and we've had 100% human written content flagged as AI and AI content not flagged. None of it is accurate

I don't work in this space...but no way can this kind of detector be accurate.

[Anonymous]

Anonymous wrote:I really don’t understand why any kid needs to use AI at all to write this essay. Not Grammarly, not Claude, not anything. Just write the essay and ask another human to proofread or give feedback.

Why prefer human feedback?