This would presumably use automatic systems developed by humans, who'd also judge the ability of the systems to gather and organize accurate and relevant inform

This would presumably use automatic systems developed by humans, who'd also judge the ability of the systems to gather and organize accurate and relevant information according to human standards. Doesn't that just push the biases into a less obvious domain?