4 ms·

> Human scale surveillance is okay. Because the surveillance is carried out by a human on the scene, it requires the surveillers to decide who should be surveil

> Human scale surveillance is okay. Because the surveillance is carried out by a human on the scene, it requires the surveillers to decide who should be surveilled, and prevents or eliminates blanket surveillance. But what we're describing is not a human using his judgment to select probable suspects for further investigation. It's a guy parked pointing a camera at all license plates and feeding them en masse to a computer for automated processing. It is not functionally different than the hypothetical robot you described, only this happens to be a robot with an organic component.

simoncion 11y ago
> It is not functionally different than the hypothetical robot you described, only this happens to be a robot with an organic component. Two things. 1) You appear to have missed this in my original comment: > It's very important that we protect the right to notice and record things that happen in public. I feel that it's equally important that we work to limit the damage that an entity can do when they have the resources available to record such things on a large scale. 2) It's obvious that one can very cost-effectively blanket any urban or semi-urban area in ALPR cameras. When you do this, you gain the ability to track any vehicle in that area forever. For areas of any appreciable size, no police force has the personnel on hand to achieve the same level of coverage with human officers. This means that -if you ban robot officers- you dramatically limit the parts of one's trips that the police can remember. It's the difference between a passive attacker that controls a few nodes in the Tor network, and a passive attacker that controls most of the nodes in the Tor network. In the former situation, that attacker may be able to use traffic analysis to unmask a few Tor users. In the latter situation, they will be able to unmask very many Tor users. [0] Make sense? [0] This is -as the Tor people have always said- not something that Tor protects against.
- hellbanner 11y ago
  Ok -- I'm following your argument now. I'm surprised that you think 100% blanket surveillance is not ok, but partial surveillance -- limited only by its expensiveness -- is OK. Because what's most likely to happen is that the police choose to surveil an area for political reasons (like racial segregation), and are thus attacking a specific group. Yes, by limiting robot police & robot cameras you make this difficult for police by limiting their memory, but mass-surveillance is still happening! This, myself & the other commenter (I believe) find this unacceptable. I'm arguing that police shouldn't be mass surveilling at all -- with robots, helmet cams that track license plates or by individual human officers selecting a zone to surveil. Or that if this surveillance happens -- it becomes public knowledge. Let every citizen view the live feeds around them, perform their own tracking if they wish, to keep the playing field open. The Tor project doesn't protect against isolated attacks (surveillance), but would it not be nice if it could? I think we have a difference of opinion here so I'm curious -- what are valid reasons from your viewpoint in which a human officer should select an area to surveil?
  - simoncion 11y ago
    > Because what's most likely to happen is that the police choose to surveil an area for political reasons (like racial segregation), and are thus attacking a specific group. That has nothing to do with police surveillance. That has everything to do with crooked cops. It's a problem that's worth talking about, but it's completely unrelated to a conversation about limiting the damage that can be done with tireless, incredibly cheap robot police officers. > ...but mass-surveillance is still happening! What I'm describing as the ideal situation isn't mass surveillance. If you declare all surveillance that gets fed back into a filing cabinet (computerized or otherwise) to be mass surveillance, you make it impossible to distinguish between a single guy on a stakeout and an organization that records and stores for the rest of eternity every single phone call in a nation, the exterior of every single piece of snail mail in a nation, and the movement of every single automobile in every major and most minor metropolitan areas in that nation. Please don't do that. > The Tor project doesn't protect against isolated attacks (surveillance), but would it not be nice if it could? If you could get it for free, it would be nice. It's a hard problem that requires some very serious tradeoffs. There is no way around those tradeoffs. Tor -as is- is very useful. If it protected against a passive adversary that controlled most of the nodes in the network, nearly noone would use it because it would be too slow. > ...what are valid reasons from your viewpoint in which a human officer should select an area to surveil? I mean, the obvious stuff? * As part of a routine crime-prevention patrol. (E.g. an officer walking a regular beat.) * As part of an ongoing investigation into one or more people who are -at a minimum- suspected of being involved in illegal activity. Let's skip the conversation about what should or shouldn't be illegal. While that's a useful conversation to have, it's not in-scope for a conversation about limiting the damage done by robot policing.

simoncion 11y ago

> It is not functionally different than the hypothetical robot you described, only this happens to be a robot with an organic component. Two things. 1) You appear to have missed this in my original comment: > It's very important that we protect the right to notice and record things that happen in public. I feel that it's equally important that we work to limit the damage that an entity can do when they have the resources available to record such things on a large scale. 2) It's obvious that one can very cost-effectively blanket any urban or semi-urban area in ALPR cameras. When you do this, you gain the ability to track any vehicle in that area forever. For areas of any appreciable size, no police force has the personnel on hand to achieve the same level of coverage with human officers. This means that -if you ban robot officers- you dramatically limit the parts of one's trips that the police can remember. It's the difference between a passive attacker that controls a few nodes in the Tor network, and a passive attacker that controls most of the nodes in the Tor network. In the former situation, that attacker may be able to use traffic analysis to unmask a few Tor users. In the latter situation, they will be able to unmask very many Tor users. [0] Make sense? [0] This is -as the Tor people have always said- not something that Tor protects against.

hellbanner 11y ago
Ok -- I'm following your argument now. I'm surprised that you think 100% blanket surveillance is not ok, but partial surveillance -- limited only by its expensiveness -- is OK. Because what's most likely to happen is that the police choose to surveil an area for political reasons (like racial segregation), and are thus attacking a specific group. Yes, by limiting robot police & robot cameras you make this difficult for police by limiting their memory, but mass-surveillance is still happening! This, myself & the other commenter (I believe) find this unacceptable. I'm arguing that police shouldn't be mass surveilling at all -- with robots, helmet cams that track license plates or by individual human officers selecting a zone to surveil. Or that if this surveillance happens -- it becomes public knowledge. Let every citizen view the live feeds around them, perform their own tracking if they wish, to keep the playing field open. The Tor project doesn't protect against isolated attacks (surveillance), but would it not be nice if it could? I think we have a difference of opinion here so I'm curious -- what are valid reasons from your viewpoint in which a human officer should select an area to surveil?
- simoncion 11y ago
  > Because what's most likely to happen is that the police choose to surveil an area for political reasons (like racial segregation), and are thus attacking a specific group. That has nothing to do with police surveillance. That has everything to do with crooked cops. It's a problem that's worth talking about, but it's completely unrelated to a conversation about limiting the damage that can be done with tireless, incredibly cheap robot police officers. > ...but mass-surveillance is still happening! What I'm describing as the ideal situation isn't mass surveillance. If you declare all surveillance that gets fed back into a filing cabinet (computerized or otherwise) to be mass surveillance, you make it impossible to distinguish between a single guy on a stakeout and an organization that records and stores for the rest of eternity every single phone call in a nation, the exterior of every single piece of snail mail in a nation, and the movement of every single automobile in every major and most minor metropolitan areas in that nation. Please don't do that. > The Tor project doesn't protect against isolated attacks (surveillance), but would it not be nice if it could? If you could get it for free, it would be nice. It's a hard problem that requires some very serious tradeoffs. There is no way around those tradeoffs. Tor -as is- is very useful. If it protected against a passive adversary that controlled most of the nodes in the network, nearly noone would use it because it would be too slow. > ...what are valid reasons from your viewpoint in which a human officer should select an area to surveil? I mean, the obvious stuff? * As part of a routine crime-prevention patrol. (E.g. an officer walking a regular beat.) * As part of an ongoing investigation into one or more people who are -at a minimum- suspected of being involved in illegal activity. Let's skip the conversation about what should or shouldn't be illegal. While that's a useful conversation to have, it's not in-scope for a conversation about limiting the damage done by robot policing.