There are many examples of codes of ethics which sound amazing, but which in practice, have little or even no impact. The authority of codes of ethics derives from different sources. They may be backed up with professional sanctions, for example, or by outside pressure from consumer groups. There is a problem with many instances of AI that there may be no corresponding professional body, since it is such a diverse field. In this way, AI has a harder problem than other professional fields such as medicine or engineering (although of course many working in AI are members of professional engineering bodies, such as the IEEE.) Codes of ethics which are aspirational and general, such as the Asilomar Principles, or the EPSRC Principles of Robotics, may act as general benchmarks but may have little impact on those who simply choose to ignore them.

Such issues are discussed at greater length in the book, Towards a Code of Ethics for Artificial Intelligence.

We would like to extend our thanks to the Future of Life Institute for their generous support of this project.