What happens when you don't attempt to correct for bias in your AIs inputs:
https://www.theverge.com/2016/3/24/11297050/tay-microsoft-ch...
People like to think that ML/AI can only produce truth, while in actuality it only reflects back what it is shown.