Defense Innovation Board Principles
Defense Innovation Board Principles
On Feb 24 2020 the US Department of Defense adopted a set of Ethical Principles for the responsible use of AI. In 2019, Seth Lazar was invited to an expert panel to provide feedback on candidate principles, and submitted a policy brief to the Defense Innovation Board, which developed the principles. In September 2019, Lazar provided substantive comments on a draft of the principles to the lead author, Dr. Heather Roff (Johns Hopkins University Applied Physics Lab).
Here’s an excerpt of Lazar’s initial submission:
Recent months have seen organisations from small corporations, to national and regional governments, and multinational tech companies, putting forward principles to guide their adoption and use of artificial intelligence systems. A number of themes have emerged, to the extent that each new set of principles tends to be a subset of the last. Most lists of principles confusingly include general goals that should be aimed at in any endeavours of a given organisation, alongside very specific issues raised by AI in particular. I welcome the DIB’s approach of not simply rehashing existing policy guidelines, but taking time to think seriously about what makes AI different, and why we need new principles to govern it. I would encourage them to spend at least as much time thinking about what makes Defense different from the other organisations that have set out these policy statements. These differences, I think, make the principles around which other organisations have coalesced much less useful for Defense. I’ll begin this comment by asking first what makes AI different, then what makes Defense different. I’ll then ask what the goal is, of developing a set of AI policy principles. Finally I’ll consider which kinds of principles we can propose. In particular, I will argue that some of the principles that so frequently appear in other organisations’ lists should apply quite differently, if at all, to Defense.