What's particularly frustrating, I think, is that there are all these problems in how someone might apply AI.

But, then, you get some shallow take like "more white people than black people might appear with a very specific search term".

Or "it might generate something offensive".

And then, these actual problems get ignored.