Reading between the lines: - the researchers noticed it and weren't quite comfortable with it - getting rid of most of the islamophobic input material isn't s

Reading between the lines: - the researchers noticed it and weren't quite comfortable with it - getting rid of most of the islamophobic input material isn't simple, so it's widespread, not a niche - the journalists who wrote this article knew that And still, the text somehow blames AI rather than the people who wrote all the source material. As it asks: "How can the AI community fix the bias in its systems?"