Re: But TensorFlow is Open Source!
Czrly, what you bring up is really at the heart of what I think the ethics of this are. Not the bit about working on open source because it can be abused - that kind of thinking leads to stagnation as anything can be re-purposed to accomplish goals other than what was originally intended. The real ethical issues as I see them are 1) when is it ethical to develop new weapons and 2) when is it ethical to use them?
The article brings up the idea that ethics are tied to risk analysis and that not enough has been done, but that is just a matter of spending the time and doing the analysis and perhaps implementing and proving failsafes. Once we are past that, we are still stuck with the above to questions.
My feeling is that AI used by a nation for political or military (what difference, really?) goal should be governed by the same rules as any other use or threat of force, but that is just my simple opinion. The first question as to when it is ethical to create a new weapon is much more complex and I don't know where to begin on it.