This is exactly where users of English as second language are being accused of cheating -- we didn't grew with the live language, but learnt from movies, classic books, and in school (the luckiest ones).
We use rare or uncommon words because of how we learned and were taught. Weaponising it against us is not just a prejudice, it's idiocy.
You're postulating using a metric that shows how much someone deviates from the bog standard, and that will also discriminate against the smart, homegrown erudites.
I’m referencing a paper I saw in passing multiple years ago, so forgive me if I didn’t elaborate the exact algorithm. The LLM varies its word selection in a patterned way, eg most likely word, 2nd most, 1st, 2nd, and so on. It’s statistically impossible for an esl person to happen to do this on accident.
I remember when my parents sent me to live with my grandparents in India for a bit, all the English language books available were older books, mostly British authors. I think the newest book I read that summer that wasn't a math book was Through the Looking Glass.
We use rare or uncommon words because of how we learned and were taught. Weaponising it against us is not just a prejudice, it's idiocy.
You're postulating using a metric that shows how much someone deviates from the bog standard, and that will also discriminate against the smart, homegrown erudites.
This approach is utterly flawed.