The whole world can see our edit summaries.
They are not just for the benefit of reviewers. They're not even just for reviewers and others who might question if an edit should be kept or partially reverted / completely rolled back. They are for anyone reading the revision history.
Edit summaries appear in every post's edit history, "for all time." They are viewable by anyone who can view the post. Edit histories are not indexed by search engines, but any person can look at them.
For example, I improved the appearance of an ordered list in this answer with the edit summary
improved formatting. Now suppose someone clicks the link to view the post's revision history:
Then they get this page, which shows:
Edit summaries are useful even when an edit doesn't require approval.
This does not necessarily mean that submitting a blank edit summary is universally bad. Sometimes an edit is simple, the justification is easy and non-controversial, and it speaks for itself more clearly than a summary. In these situations, it is acceptable to submit a blank edit summary, because the system automatically adds syntax-based edit summaries when none is provided. (Like
edited tags, and
removed 13 characters from body.)
However, as Oli says, it is never wrong to provide an edit summary. (Unless the edit summary is wrong. Or provides no information. Like the
improved post edit summaries I complained about there.)