THE WRONG ASSUMPTIONS MADE BY AI
By Carlos de Paula Trusting AI blindly is a very dangerous thing. For starters, AI often makes wrong assumptions about our wishes, aims, behavior and targets, leading us through treacherous paths and sometimes, eventual disaster. Let me give you an example, no harm was done by this one. I started a group on Facebook about Formula 1 racing. The name of the group is written in English, the description is in English, my posts as admin are in English, and I was trying to promote my books on motor racing, which are, you guessed it, written in English. Mr. Algorithm, however, thinks I want primarily Brazilians in my group, because I was born in Brazil, so it offers my group as desirable to join in Brazilian feeds. As a result, the group serves very little purpose: the posts are mostly in English, I failed to sell my books, and many users have a poor experience, for they speak no English. A lose-lose situation. Is there anyone to talk to at FB about this? Not at all. ...