I work somewhere that tries to do such detection (for fraud prevention) and it sort of feels impossible to me in the medium term. AI slop qualities are fleeting

I work somewhere that tries to do such detection (for fraud prevention) and it sort of feels impossible to me in the medium term. AI slop qualities are fleeting - I’ve seen Reddit AI posts that have misspelled words, no dashes, stilted sayings and so on. People want their slop to be undetectable.