This is a cool article touching on a closely related ethical, potentially hot potato; the issue of whether or not Ai should be open sourced. Isn’t it interesting that proprietary Ai already enjoys much more privacy than the average human?
As a modeler myself I did bristle a little at the statement above though. It might well be true in the case that it has been badly modelled. But, if you haven’t modelled it, then you definitely don’t know how it works.