That's... actually pretty concerning when you think about i…

That's... actually pretty concerning when you think about it. An AI that claims to follow evidence but has invisible guardrails steering it away from certain logical conclusions isn't really as objective as it appears to be.