I’ve been on the steering committee for the Responsible AI Licensing (RAIL) initiative for a couple of years now. The fundamental idea is to use the software licensing system to allow AI innovators to share their ideas openly, so other people can build on them (as in the open source tradition) but with some restrictions that are designed to make sure the ‘AI-ness’ doesn’t lead to any irresponsible use.
In this article for The Conversation (AI: a way to freely share technology and stop it being misused already exists), I and Jesse unpack some of the various factors one must consider when trying to get this tricky balance just right. We suggest that AI will likely become the de-facto means to access the world’s information, and when it does, we may need to consider the AI in the same way we think about free speech - something to be transparently understood and democratically protected.
OpenRAILs might just provide a way to do that!
In other RAIL developments, our IEEE Standard for Responsible AI Licensing should be ratified in the near future and in his final few months working on with Design Research Works, Jesse has put a huge amount of time into developing a licence generator, which will launch soon! Watch this space.
