There has long been the procedure in place where add-ons are submitted by the developer for testing.
But according to the guidelines, developers didn't have to submit them for review and that is where the problem was.
And again, with this new policy, this will be an automated process and only if that fails, will "people" get involved so until then, bias "should" not be an issue. It is all about ensuring the add-on does nothing without the user's awareness/consent and/or without being able to undo/uninstall. If the automated process cannot verify those requirements, then humans will get involved.
At least that is how Mozilla claims the process will work.
As far as the review team being biased, I guess only time will tell. I sure hope not. I also hope they make no modifications to the code, other than signing it. If they deem it necessary to make further changes, they need to reject the submission and return it to the developer rather than modifying it and then signing it.
As far as Moonchild's complaint, that is way beyond my level of expertise, but if Mozilla is only "signing" the extension, then even though that is technically modifying the code, I don't see the problem if it, in no way, modifies the function of the add-on.
To be sure, I respect the developer's rights to develop their own code, but when it comes to consumer security, the consumer's right to security trumps the developer's rights.
Freedom is NOT free and unfortunately, it is the honest people who must bear the burden and costs of those who would be dishonest.