Document Type

Article

Publication Date

8-23-2023

Abstract

This somewhat tongue-in-cheek post discusses biases in AI systems. Noting that AI bots need to be “trained,” this post suggests that untrained mediator bots may spew out unwanted interventions such as providing undesired evaluations of BATNA values – or failing to provide desired evaluations. So mediators probably will need to co-mediate with their bots for a while to observe and correct its biases. Ironically, bots may produce language that normal humans understand much better than the confusing jargon we habitually use. So the mediator bots may need to train human mediators.

Share

COinS