> There should be a dashboard indicator or toggle to visually warn when the bot is just uncritically agreeing I would be very surprised if it were possible to

> There should be a dashboard indicator or toggle to visually warn when the bot is just uncritically agreeing I would be very surprised if it were possible to reliably detect this. In fact, I'm not certain it's a distinction which can meaningfully be made.