This is what’s known as Algorithmic bias.
It’s when AI produces discriminatory outcomes as a result of being trained on data that is biased which results in these AI systems replicating and reinforcing those biases.
That very well could be why Google Doc’s spellcheck is regurgitating this kind of bigotry because it’s literally what it was trained to do.
Side note this is why BAME folk are so concerned about the rise of facial recognition for criminal investigations. Cause not only is facial recognition less accurate for us than it is for white folk (cause they make up the majority of that trained data) but also because the police systems are institutionally racist.
Which is being reflected and perpetuated by these AI systems I mean people are already being misidentified and wrongfully arrested because of this.