AI learned what it "knows" from us, the internet, books presumably all human sources.

The content described as undesirable here is freely available on the internet where AI learned it.

We accept this content is available otherwise, but if AI repeats, it that is bad?

In some ways AI seems like the whipping boy for human issues.