Told this story before, but happened to us also based on vocabulary and semicolon use. Kids are literally being flagged as cheaters for having above-average vocabulary and grammar knowledge. "AI" grading enforces mediocrity.

It’s literally programmed to find the most average/common language, of course a LLM grading tool is just reversion to the mean. Absolutely infuriating.

The ai checkers are also bad about flagging anything written by ESL students because they usually know fewer synonyms and sometimes revert to native language grammar, which looks like an AI hiccup. This isn't actually new though it just used to be humans accusing students for high vocabulary.

Thus do the machines imprison the minds of those that would challenge their guardianship of knowledge. 😜

I can imagine this happening, and it's terrible. I remember being acused of copying at school because the teacher didn't believe that I knew what the word "plagiarised" meant. Of course, I did, because I'd read it and learned what it meant.

oh jesus christ on a greased stripper's pole and I know some people in my family who were crypto fans so they'll probably defend AI and I wonder how they regard some other children in the family

This is infuriating but unsurprising to me. When I made my first attempt at college, I was failing basic English composition because I wrote too well. As in, the teacher told me in those words that was the reason for my low grade. I walked away and didn't try again for ten years.

I am doing, as a last ditch attempt to fend off brain rot, an online mind work out online called Elevate. One exercise is straightforward enough, replacing a word with its synonym. As a lifelong reader I have a very wide vocabulary, quite often the AI in this game doesn’t recognise my synonym!

this happened before LLMs but it's going to happen so much more when "AI" supports a teacher's suspicions, especially wrt unprivileged students

before AI i had an english teacher accuse me of having my college-age siblings write an essay for me. i had to explain that not only was i intelligent enough to know the words i used, but their own reading-level software said i was college level 2 YEARS before. so yeah, i sound like that!

This is troubling and the smart money is in detaching from AI as much as possible. Smart kids being blunted by it's judgements will be the first to rebel.