Ethics by Agreement in Multi-agent Software Systems
Interesting paper I found while revising my most recent article "Should we embed machines with ethics that we know to be good (ethics-by-design), or should we repose our faith in a ethics-making method that leads to ethics emerging by agreement in a society of machines?" As the authors suggest, the former sounds like the best option, but in a world with a multiplicity not only of ethical perspectives, but also a multiplicity of interacting AI systems, it's impossible to achieve. So, better to consider how AIs expressing the various perspectives can come to agreement on ethical questions.
Today: 3 Total: 453