Delete comment from: Captain Capitalism
As I understand it, the AI was 'gamed' by a dedicated group of internet trolls who thought it would be funny if they exposed an 'innocent' AI to the "horrors of the internets". They were apparently very successful, and since it was done to an AI, not a person (as in a being with personhood, not a human being, AIs do not have personhood or even sentience), I actually find it a little funny myself.
However, don't read too much into it. It was probably even easier to game the AI to be, do, or say those terrible things than it is to game the Google search engine algorithm to rank a website higher (Search Engine Optimization), and there's an entire industry dedicated to that goal. The only way to see how an AI would learn things from the internet, the first step would be NOT TELLING THE INTERNET THAT YOU ARE AN AI THAT INCORPORATES WHAT YOU HEAR INTO YOUR KNOWLEDGE BASE!
Mar 26, 2016, 5:40:11 AM
Posted to Why Too Much Social Media is Toxic