AI focused DiploTechy writer of fiction & non-fiction about the future of tech & humanity. Author of Broad Horizons http://amzn.to/1UxH4aE Opinions mine not USG
Voting up good critiques and flagging inappropriate content as spam will also be managed through an API. Exactly how this will be managed will require iterative experimentation during the building and expansion phase because this is the most mission critical part of the whole system, and most prone to manipulation.
(b) Where the systems are functioning as expected, and where the expectations assume imperfections, the victims ought to be compensated along the model of no-fault insurance: The families of the 5,000 people killed in car crashes ought to be compensated according to guidelines that try to be as fair as possible.
…ividual users. For example, the passengers in an AV might want to optimize a trip for scenic value. Moving that “slider” up — someday possibly a slider in a digital control panel — will automatically move some of the others down: the trip is likely to take longer and consume more energy. The limits imposed on the users’ ability to adjust the sliders will be determined by those who are …
So perhaps we should take a different approach to how we’ll settle these issues. Perhaps the “we” should not be the commercial entities that build the AI but the systems we already have in place for making decisions that affect public welfare. Perhaps the decisions should start with broad goals and be refined for exceptions and exemptions the way we refine social policies and laws. Perhaps we should accept that AI systems are going to make decisions based on what they’ve been optimized for, because that’s how and why we build them. Perhaps we should be governing their optimizations.