4 ms·

CCTV security aimed at the register will capture your face, perhaps a parking lot camera will grab your license plate... While the chances of someone using tha

CCTV security aimed at the register will capture your face, perhaps a parking lot camera will grab your license plate... While the chances of someone using that information to track you down is pretty slim, it is still possible.

InclinedPlane 17y ago
The chances of someone using that information depends very critically on its nature and accessibility. Currently security cam footage tends to be localized, kept on-site only on physical media (tape or hd). It also tends to have a low shelf life (days) extremely low quality (think youtube at launch) and very little meta data (camera and date/time). But one can imagine a near future where a lot of these aspects are different. Where year or multi-year long buffers of footage are archived in a network accessible manner. Where technological advances lead to massive increases in fidelity (from 200 lines of blurry interlaced video on an overused magnetic tape is replaced by 60fps 1080p). Where extensive pre-processing generates massive amounts of meta-data. Right now most CCTV footage is just raw video. But imagine a system that kept track of persons, and collated appearances on different camera feeds per store visit. It's easy to imagine how such a system might work by, for example, comparing physical location and perhaps clothing color to connect appearances on different cameras. Then, when one reviews camera footage it's not merely a raw dump of video from different camera locations instead it's a list of visits to the store. Imagine a table listing the time of store entrance, the time of store exit, the departments visited and for how long, the time of checkout, the list of items purchased and method of payment, a high resolution headshot, and even perhaps the name (depending on method of purchase and whether they used a loyalty card). It's trivial to imagine how this technology could be expanded from one single visit to linking every visit to the store (using loyalty card, cc/debit card, or automated facial/body recognition as a means for correlation), and from there it's easy to imagine mining this data for further metadata (typical times and durations for visiting the store, typical items purchased, typical store sections visited, highlights on atypical purchasing or other activities, etc.) One can imagine a lot more metadata being extracted with more advanced techniques. Such as determination of ethnicity, income, job, friends and family, typical schedules, etc. based on sharing information from every business (e.g. not just the grocery store but also the parking garage, the restaurant, the hardware store, the DMV, google, etc.) None of this is particularly technologically difficult, it's just expensive and there's no reason for most businesses to do it. But if you look at, say, the recent investigations in Dubai related to the Mahmoud al-Mabhouh killing it's easy to see that a lot of this technology is already in use in some places. The logical limits of the technology are frightening. One might imagine that such systems could automatically track such things as your mood, job status, and interpersonal relationships with a high degree of fidelity. It could roughly determine your mood and emotional state through your spending, travel, entertainment habits, and socializing. For example, are you uncharacteristically not socializing (online or offline) with your friends and instead watching a movie to cheer you up along with a pint of your favorite ice cream? Everyone has their habits and patterns, with enough surveillance and sophisticated enough processing it's possible to tease out these patterns in a completely automated fashion. Now imagine if you happened to come under active, rather than merely passive, surveillance with such technology in place. Your phone and your internet connection could be tapped, and this information (your email, your text messaging, your voice conversations, your private irc messages, your personal credit card purchases) could be added to this database. It doesn't take much imagination to think about the amazing depth and breadth of information that could be extracted from this data given sufficient technological resources, a talented development team, and skilled investigators.

InclinedPlane 17y ago

The chances of someone using that information depends very critically on its nature and accessibility. Currently security cam footage tends to be localized, kept on-site only on physical media (tape or hd). It also tends to have a low shelf life (days) extremely low quality (think youtube at launch) and very little meta data (camera and date/time). But one can imagine a near future where a lot of these aspects are different. Where year or multi-year long buffers of footage are archived in a network accessible manner. Where technological advances lead to massive increases in fidelity (from 200 lines of blurry interlaced video on an overused magnetic tape is replaced by 60fps 1080p). Where extensive pre-processing generates massive amounts of meta-data. Right now most CCTV footage is just raw video. But imagine a system that kept track of persons, and collated appearances on different camera feeds per store visit. It's easy to imagine how such a system might work by, for example, comparing physical location and perhaps clothing color to connect appearances on different cameras. Then, when one reviews camera footage it's not merely a raw dump of video from different camera locations instead it's a list of visits to the store. Imagine a table listing the time of store entrance, the time of store exit, the departments visited and for how long, the time of checkout, the list of items purchased and method of payment, a high resolution headshot, and even perhaps the name (depending on method of purchase and whether they used a loyalty card). It's trivial to imagine how this technology could be expanded from one single visit to linking every visit to the store (using loyalty card, cc/debit card, or automated facial/body recognition as a means for correlation), and from there it's easy to imagine mining this data for further metadata (typical times and durations for visiting the store, typical items purchased, typical store sections visited, highlights on atypical purchasing or other activities, etc.) One can imagine a lot more metadata being extracted with more advanced techniques. Such as determination of ethnicity, income, job, friends and family, typical schedules, etc. based on sharing information from every business (e.g. not just the grocery store but also the parking garage, the restaurant, the hardware store, the DMV, google, etc.) None of this is particularly technologically difficult, it's just expensive and there's no reason for most businesses to do it. But if you look at, say, the recent investigations in Dubai related to the Mahmoud al-Mabhouh killing it's easy to see that a lot of this technology is already in use in some places. The logical limits of the technology are frightening. One might imagine that such systems could automatically track such things as your mood, job status, and interpersonal relationships with a high degree of fidelity. It could roughly determine your mood and emotional state through your spending, travel, entertainment habits, and socializing. For example, are you uncharacteristically not socializing (online or offline) with your friends and instead watching a movie to cheer you up along with a pint of your favorite ice cream? Everyone has their habits and patterns, with enough surveillance and sophisticated enough processing it's possible to tease out these patterns in a completely automated fashion. Now imagine if you happened to come under active, rather than merely passive, surveillance with such technology in place. Your phone and your internet connection could be tapped, and this information (your email, your text messaging, your voice conversations, your private irc messages, your personal credit card purchases) could be added to this database. It doesn't take much imagination to think about the amazing depth and breadth of information that could be extracted from this data given sufficient technological resources, a talented development team, and skilled investigators.