3 ms·

I realize I am going off topic here. But > It has no real impact on anything beyond scale, which may make a practical difference but makes no fundamental diffe

I realize I am going off topic here. But > It has no real impact on anything beyond scale, which may make a practical difference but makes no fundamental difference to any arguments about surveillance. Scale makes a huge difference: The obvious example, discussing this on HN, is computers. Fundamentally, the massive network of computers making up the internet is no different from the original Zuse computer - both are just number crunching machines. But in the same way that AI surveillance will allow you to control everyone, not just a few individuals, the difference the scale makes is just so huge that the two aren't even remotely comparable.

thu2111 7y ago
Yes, it makes a practical difference but in an academic debate setting, arguments for or against pervasive surveillance are usually scale invariant. If it wasn't you could end up with arguments like this one: "The USSR compiling detailed files on millions of citizens was not unethical because it was small scale, but Google compiling detailed files on billions of citizens is, because it's large scale." or alternatively "Surveilling terrorists is fine because there are very few of them, but CCTV to catch thieves in public places isn't because that's much larger scale." These would be very odd and brittle arguments though. In fact I've never seen anyone make an argument like that. The tolerability of surveillance depends very much on context like who is doing it to whom and why, with scale being a pretty small aspect. When people debate surveillance it's always in the context of identities and purposes. But the notion that AI+surveillence is special ignores questions of who's the watcher, who's the watched and what the purpose of it all is to focus on the least important part of all: concrete cost to the watchers. After all communist regimes already demonstrated that you can scale mass surveillance up to huge levels without needing any technology at all, just ideology. That's why I'm suggesting debates about AI in left-wing media are actually not about AI at all, that's just a way to restart the discussion about standard left-wing ideas and topics, like capitalism or sometimes identity politics. This article is an exemplar: https://www.theguardian.com/commentisfree/2019/aug/13/the-guardian-view-on-surveillance-big-brother-is-not-the-only-watcher-now https://www.theguardian.com/commentisfree/2019/aug/13/the-gu... In fairness, it does start by recognising that "big brother" is the primary user. But the thrust of the article is about private (ab)usage. It ends by saying: Imagine a surveillance system that falsely identifies a black man from his face and then just as falsely attributes to him aggressive intent from his expression and the way he is standing.
- read_if_gay_ 7y ago
  Good points. Thanks.

thu2111 7y ago

Yes, it makes a practical difference but in an academic debate setting, arguments for or against pervasive surveillance are usually scale invariant. If it wasn't you could end up with arguments like this one: "The USSR compiling detailed files on millions of citizens was not unethical because it was small scale, but Google compiling detailed files on billions of citizens is, because it's large scale." or alternatively "Surveilling terrorists is fine because there are very few of them, but CCTV to catch thieves in public places isn't because that's much larger scale." These would be very odd and brittle arguments though. In fact I've never seen anyone make an argument like that. The tolerability of surveillance depends very much on context like who is doing it to whom and why, with scale being a pretty small aspect. When people debate surveillance it's always in the context of identities and purposes. But the notion that AI+surveillence is special ignores questions of who's the watcher, who's the watched and what the purpose of it all is to focus on the least important part of all: concrete cost to the watchers. After all communist regimes already demonstrated that you can scale mass surveillance up to huge levels without needing any technology at all, just ideology. That's why I'm suggesting debates about AI in left-wing media are actually not about AI at all, that's just a way to restart the discussion about standard left-wing ideas and topics, like capitalism or sometimes identity politics. This article is an exemplar: https://www.theguardian.com/commentisfree/2019/aug/13/the-guardian-view-on-surveillance-big-brother-is-not-the-only-watcher-now https://www.theguardian.com/commentisfree/2019/aug/13/the-gu... In fairness, it does start by recognising that "big brother" is the primary user. But the thrust of the article is about private (ab)usage. It ends by saying: Imagine a surveillance system that falsely identifies a black man from his face and then just as falsely attributes to him aggressive intent from his expression and the way he is standing.

read_if_gay_ 7y ago
Good points. Thanks.