3 ms·
It might take weeks for someone to identify you in real life based on a grainy still image from a camera. And then one would still have to correlate among peopl
by nnfy 9y ago
It might take weeks for someone to identify you in real life based on a grainy still image from a camera. And then one would still have to correlate among people, other cameras, and locations to learn anything about you without surveillance.
With the internet, because of data aggregation, I can go on piple and find all of that about you and more in 2 minutes.
The power with which one may invade privacy over the internet is unprecedented. Let's say you frequent a coffee shop commonly visited by republican groups. In just a few minutes, in theory, someone could find out about every single time you visited, and use that against you. Replace republican coffee shop with sex shop, doctor, psychologist, democrat gathering...some things are better left hidden, because people are terrible.
- stephengillie 9y agoSo it's bad because it's faster? Or easier to automate?
- dredmorbius 9y agoScale, speed, and costs all matter. The Jevons paradox notes that an increased efficiency of some process or system increases the total utilisation. In particular, it makes previously nonviable applications viable. Postal junk mail, telemarketing calls, email spam, popover/popunder ads, malware, robocalls, fake news, chatbots, and more, are all responses of previously nonviable applications becoming viable. The underlying limitation seems to be the scope of attention of attackers based on private / personal data, and/or of the systems in which they operate. The role of AI in extending that institutional bandwidth ... strikes me as rather frightening. There've been previous discussion of similar topics which point to the prospect of, say, automated lawsuit filing, or debt collections (already a problem), or more. The prospect of some trained algo running over deep, rich data, seeking arbitrage opportunities, strikes me as undesireable.