Does it not concern you in the slightest, that when they've ended up speaking to higher and higher up people in the various organisations that they literally don't know why the algorithm decided the way it did? No one they've spoken to can figure it out, which means they themselves can't also rule out discrimination happening.

A lot of the technology we produce is fundamentally biased by our own perceptions, and in ML case most especially by the input data. The danger is a self-perpetuation of already established systemic bias. Given access to good credit is key towards economic mobility that can have massive ongoing implications. The fact that they're saying "It's just what the algorithm produces" should be triggering huge alarm bells, regardless of the veracity of any of his other claims.