The term 'Autonomous and Intelligent Systems' (A/IS) stands for what we would commonly call AI and this is an extensive effort to describes the ethics of A/IS system design. This document (266 page PDF) will take some time to read. Take the time. I applaud the good intentions of the IEEE here but just as you wouldn't want a philosopher to design a bridge, a document that reads to me a bit like "Ethics by Engineers" feels wrong. For example, consider one of the "goals" of the document: "Prioritize metrics of well-being in their design and use." It's a basic error to substitute 'metrics' for outcomes, but they make it here. The "concerns" include automated weapons, safety, affective computing and mixed reality. OK. But where are the concerns about stereotyping and misrepresentation, which are the actual problems today? But beyond any such mundane concerns is "the seminal question of determining the key performance indicators (KPIs) of their success once introduced into society." Sigh. The IEEE website will make you fill in a form to access this document, so I uploaded a copy to my own site so you can access it directly. For more ethics as designed by engineers, see the IEEE Ethics in Action website.
Today: 5 Total: 98 [Direct link] [Share]
