I spent a few hours working through the discussion in this document. The researchers have selected and analyzed 36 documents that take a 'principled' approach to ethics in artificial intelligence - that is, they identify or articulate some set of concerns or principles of actions to define ethical AI. While the document seems at first to document a convergence on themes, it becomes clear on closer reading that "there's a wide and thorny gap between the articulation of these high-level concepts and their actual achievement in the real world." I woul also point out that because of the nature of the documents selected - corporate and institutional studies - there is a whole swath of opinion not represented. The document appears to be mostly a means to structure the debate, and it certainly does that (see the diagram here), but I don't think an analysis of sets of principles is the way to approach such a structuring.
Today: 4 Total: 96 [Direct link] [Share]
