3 ms·
I agree, and your comment hints at a degree of optimism. There are individual cases of privacy breach and institutional ones. The primary processual means of
by sixdimensional 8y ago
I agree, and your comment hints at a degree of optimism. There are individual cases of privacy breach and institutional ones.
The primary processual means of protection we've come up with appears to be policy and regulation. Cultures differ so different policy and regulation in different places.
However, I'm seeing a trend of justifying privacy intrusion through the argument that "only the machine learning algorithm sees it". For any purpose, governmental or not, algorithmic or human, the breach of privacy is a breach.
I guess naturally we will likely find some kind of balance... or maybe not... but I do still believe we will be forced to confront the fact that we will see and know more of each other.
I'm not sure, that might be a good thing in one light, having to face each other, and the end of the concept of natural privacy on the other. Perhaps only those with the means to enforce their privacy will be able to in the future.
I think that agrees with your point that outside of certain situations it might come at a premium. Or maybe it will come at a premium regardless of environment.
I'm reminded of the physics concept "the observer effect", by observing something you change it. That change is what we must face when we are almost always being observed.
- norswap 8y agoMichel Foucault has some interesting thing to say on this, in the context of his study of the carceral system ("Discipline & Punish").