3 ms·

I have two issues with automatic surveillance. One is of course false positives. When you scan everything, even a small false positive chance will mean a LOT of

I have two issues with automatic surveillance. One is of course false positives. When you scan everything, even a small false positive chance will mean a LOT of people get swept up by false positives. Scan enough things, and soon everyone will be victims of false positives. This is compounded because people don't understand probabilities. Take DNA... if someone says "the chance of this DNA match being wrong is one in a million", people will assume the match is correct, because that seems like a really small chance of a false positive.. however, if you are matching that DNA to everyone in the United States, that one in a million ends up with over 300 matches that are wrong. Second, perfect enforcement of laws is not always ideal. If we catch everyone who breaks the law, how can we ever decide as a society to change the law? For example, if we had 100% effective anti-marijuana laws, no one could have ever tried marijuana and we would not have been able to determine that we actually don't want it outlawed. If no one can ever break a law, we have no way of testing whether a law is good or not.

seoaeu 5y ago
> Scan enough things, and soon everyone will be victims of false positives. If there is a human in the loop checking for false positives, then I don't see how it could get to this point. It surely wouldn't take hundreds of millions of false positives -- each with a non-trivial cost for manual review -- before pulling the plug.
- cortesoft 5y ago
  Yes, humans could help correct false positives... however, that doesn't mean there are no consequences for the people wrongly matched. Investigators might have to look into the person, and question them. Not sure if you have ever had to talk to the police about anything, but the situation is stressful even if you are completely innocent. Also, having your life examined, having to answer questions about why you did things, etc, can be scary.
DangitBobby 5y ago
You also must trust that it's actually automatic, kind of like police dogs. Turns out their main function is to provide "probable cause" when their handler wants a search. This opens the door to the same thing except now it's a free warrant to access any device, as long as you pay Apple to automatically "detect some CSAM" first.

seoaeu 5y ago

> Scan enough things, and soon everyone will be victims of false positives. If there is a human in the loop checking for false positives, then I don't see how it could get to this point. It surely wouldn't take hundreds of millions of false positives -- each with a non-trivial cost for manual review -- before pulling the plug.

cortesoft 5y ago
Yes, humans could help correct false positives... however, that doesn't mean there are no consequences for the people wrongly matched. Investigators might have to look into the person, and question them. Not sure if you have ever had to talk to the police about anything, but the situation is stressful even if you are completely innocent. Also, having your life examined, having to answer questions about why you did things, etc, can be scary.

DangitBobby 5y ago

You also must trust that it's actually automatic, kind of like police dogs. Turns out their main function is to provide "probable cause" when their handler wants a search. This opens the door to the same thing except now it's a free warrant to access any device, as long as you pay Apple to automatically "detect some CSAM" first.