Current AIs face the challenge of filtering bias out of the corpus they are trained on. As AI researchers have pointed out in, for example, predictive policing projects, the danger is mainly in propagating and automating previous biases.

An AI that isn't biased is likely to please Elon even less.

In my last submission I asked, Ask HN: Are leaders genuinely afraid of AI or do they have an agenda?

Here's my answer

>might be used to "flood our information channels with propaganda and untruth."

Surely that raft was flushed down the Sewer some time ago!