How to design systems that respect prerogatives of civil liberties, privacy, etc.? AG suggests five ethical principles:
Default to harmlessness. Everyware “should default to a mode that ensures their users’ safety.” It’s beyond graceful degredation, because everyware takes so much responsibility upon itself to take care of people.
Be self-disclosing. You should be able to see what systems are operating in a space, both to geeks and to people who aren’t wired up. This requires “a new universal vocabulary of signs” for everyware; and the ability to look under the hood.
Be conservative of face. Everyware should not “unnecessarily embarrass, humiliate or shame their users.” Nor should it completely dissolve the boundaries of privacy that people expect.
Be conservative of time. Don’t “introduce undue complications into ordinary operations.” Having physical equivalents of Clippy the Office Assistant would be a pain.
Be deniable. Everyware “must offer users the ability to opt out, always and at any point.” If ubicomp systems offer some functionality and benefit, opting out should just turn those off. (How do you opt out of being photographed by surveillance cameras?)
I think these are fine principles for private-private relationships. I do not think they work for individual-state relationships. Also they seem to be a bit devoid of the economic relations in favor of social relations. Time for instance is as much an economic relation as a social relation.