Where should the line be between AI checking our decisions and AI making them?
retained text
Human judgment is biased, tired and careless, I’m not romantic about it. So I like AI as a second reader: it catches mistakes, spots patterns I miss, pushes back on my assumptions. But I keep noticing how easily the second reader becomes the author. The suggestion gets accepted, then expected, then nobody remembers who actually decided. My rough rule so far: let the machine check the work, never let it sign it. A system can be audited, but only a person can be asked “why did you do that?” and have to answer.
document fetched 2026-11-08 00:18 · canonical https://reddit.com/r/artificial/comments/1xq5iuy/where_should_the_line_be_between_ai_checking_our