[spicymapotofu]

The line that separates good psych science from bad is more subtle than word choice or tone, and often researchers who primarily do good work may present work with faulty conclusions without anyone knowing better. It has taken decades for the replication crisis to take this shape, and maneuvering of data, subjects, and results toward outcomes that don't reproduce is systematic and subtle. I'm not suggesting AI couldn't find this distinction, but what remains after is the rigorous proving of the AI conclusion - which might take just as long as if AI wasn't involved.