Fraud really thrives in moments of great social change and transition. We’re in the midst of a technological revolution. That gives con artists huge opportunities. People lose their frame of reference for what can and can’t be real.
"...We’re in the midst of a technological revolution. That gives con artists huge opportunities. People lose their frame of reference for what can and can’t be real."
So I suppose the hypothesis is that more people have defective bullshit detectors during big changes in any given field. A possible test of the hypothesis would be to look at fraud prosecutions by domain of deception (high tech stocks, property development) and by choice of victim (newly rich technical people, people new to a country &c) and see how those vary with time.
That is not obvious to most people, because the concept of customer service we understand, while we really don’t understand learning algorithms.
But technically Microsoft definitely can know that, if that were its goal. There's nothing technologically lacking to have that knowledge. Most probably it simply doesn't use that information, specifically doesn’t call random users.
They didn't say fraud doesn't exist otherwise.
I do suggest looking into crime statistics for fraud and how they're determined over time if you want a more objective, representative viewpoint.