The usecases that most people see are the Firehose of Falsehood / Engine of Sockpuppet / Election Disinfo / Author & Artist Plagiarism uses, not the quiet, “arbitrary social media platform now has ~2% Toxic content down from ~18% 5 years ago” usecases
Comments
There are syntheses (“it’ll take my job and do it badly”) but not imo enough reckoning in discourse overall with that tension
It requires prior mastery in a subject for a human to spot when the AI is incorrect. Or extensive independent verification of results