I’m not sure I agree with the article.
The article says that any algorithm designed to ban domestic terrorists would end up banning too many Republican politicians. I’m not buyin’ that. It’s easy enough to build robotic filters with artificial intelligence systems which would ban some accounts (let’s say for using the “n” word as an example) but only screen other accounts, leaving the final decision up to human review. That process, if created properly, would allow the artificial intelligence to improve its own algorithms based upon which of its original screening selections were overridden, thus gradually reducing the human role in the process as the “bots” learn what the humans want.
As I see it, the real reason is this:
Twitter is a for-profit enterprise which bases its decisions on the health of its business model. They don’t want to ban anyone unless one of two things is true: (1) the user commits or conspires to commit a crime for which Twitter could be held responsible; (2) Twitter would make more profit by banning the user than by allowing him or her to stay.
Continue reading ““There’s A Simple Reason Twitter Won’t Ban All The Nazis”” →