There is also a need to recognize that a lot of this science just doesn't work (to the extent that it is often presented). The right way to phrase a lot of the conclusions of these experts, at the state of the art, would be 'in my expert opinion, it looks like it might be their DNA/fingerprint/hair/teeth', not 'in my expert opinion it is [...]'. Because even the state of the art is often about at that level - probably closer to 75% accuracy than 99.99% like it's often treated. Especially on real world, partial, corrupted samples.

And to emphasize again, I'm talking about the state of the art without the biases being discussed. The biases only compound this problem significantly.