Hm ... apart from the fact that it could be possible to support decision making in future by AI, would you be in favour of deciding about every single flag?
I had in mind that such a 'committee' would only get active if someone complains ... but of course that's not the only thinkable option.
Not every single flag... I would think it as, "you need to bring a special case". In my view the community should be the one to flag. But on situations of power or scheme where the user is somehow imprisoned either by lack of power or because of "bugs"... this committee should be there "to make things better". They might not be able to solve all problems, and that is a consequence of who votes for the committee.
AI will mostly (or should) help with analysis of things. Giving you valuable information about what's happening without you having to go and look for everything (which is practically already impossible on STEEM).