When these techniques fail, they fail in ways that are not human. AI is simple math, there’s no room in it for ethics, or common sense, or empathy. Some of the egregious errors we will catch, but a lot of them we will not. These systems, which would require an extraordinary level of oversight to run safely, are being deployed across the planet by naive Stanford grads with almost no human supervision, for profit, and have the potential to profoundly reshape our society.

Poynter.org interviews Maciej Ceglowski: Do Facebook and Google have control of their algorithms anymore? A sobering assessment and a warning