[seabea]

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.

[none_to_remain]

The journalism around that incident seems typically silly- like my shell should be considered racist because I can type 'echo "Hitler did nothing wrong"'