Oh I hadn't considered the statistical advantage here.

You do lose out on a lot of human bias in the research process, but you also create blind errors that are hard to validate.

I know in my work there is plenty of times I run analysis and go back and manually check some entries as a sanity check - pros and cons here!