In my field (sociology) double blind review is standard. A reference "we (2018) shows" identifies the authors, so it cannot be part of the review copy. In principle, you could have a review copy that says "A and B (2018) show", and the final copy replace it with "we (2018) show", but that is rarely done. It is usually left as "A and B (2018) show".

The I and we is less of a problem in my field. In my field (sociology) double blind review is standard. A reference "we (2018) shows" identifies the authors, so it cannot be part of the review copy. In principle, you could have a review copy that says "A and B (2018) show", and the final copy replace it with "we (2018) show", but that is rarely done. It is usually left as "A and B (2018) show". Now your colleague cannot say "A (2018) shows" as that does not correctly refer to the entry in the bibliography, so it will always be "A and B (2018) show".

This is a non-intentional side-effect of this tradition, but I think a good one. I would consider all co-authors responsible for the entire paper.