John Ridgway discusses the challenges of designing autonomous Artificial Intelligence decision-making systems to be ?sufficiently safe?. As ?sufficiency? requires a value judgement in the safety-engineering domain, he suggests that it is more ?artificial wisdom? that is necessary, which will require AI to operate within a suitably encoded ethical framework. But how would such a framework be encoded, on whose value system would it be based and how much should emotion play in decision making? John explores these questions and provides some insights into tackling these issues.
John Ridgway (Wed,) studied this question.