You Look Like a Thing and I Love You by Janelle Shane

6 out of 10

Machines learn through trial and error. Typically, they start from scratch and explore a vast data landscape in search of narrow optimization parameters. The algorithm regularly allows them to exploit unforeseen opportunities, occasionally causing destructive results.  We humans should be cautious about trusting AI because it is only as smart as the data it is trained on--which is prone to be messy, biased, dubious, and outdated.