Comment by elevation

> AI [...] catching mistakes

Make sure you document why/how you determined that a previous research conclusion was a mistake -- generations after you may discover additional information and should be able to weigh the same evidence rather than simply accept (or reject) your novel conclusions.