CBC Radio's The Current had an interesting in-depth report this morning on the politics of RFID (Radio Frequency Identification Device) tags (in part 2 of the show). Canada's privacy commissioner singled out RFIDs as a cause for privacy concerns in her report to Parliament this week, so the current spoke to a few people about the issue in order to present a picture of what RFIDs are and why we should care.
RFIDs are tiny chips that can be embedded in just about any item -- including consumer goods -- and either be scanned for identifying information, or (in the case of battery-powered active RFIDs) actively transmit identifying information. The prospect of widespread or ubiquitous RFID tagging offers a lot of benefits for things like inventory management, as well as some possibilities for neat network-to-real-world hacks that appeal to the geek in some of us.
The adoption and guidelines around RFID use is a looming political issue, however, because putting specific identifiers in all the objects around us creates possibilities for personal profiling and tracking that dwarf today's concerns about the privacy implications of credit records and loyalty cards. The Current does a good job on introducing this issue -- it's worth listening to the RFID segment. (RealPlayer file; you'll need to wait through/skip through the letters segment to get to RFID.)
RFID accells at closing the gap between the real and the digital. The privacy concerns stem from the notion that the signal can be correlated to the real.
The signal itself can be encrypted, but the encrypted signal can still be correlated to the real. Instead of a signal that says a 36C bra, the signal might say 342331, which always correlates to a 36C bra.
But, the reality is that the bra is acutally given a random ID. The random ID is correlated to a particular SKU only in the SKU table. This means that 342331 would coorelate to the 36C bra this week, and after it is sold, the ID can be recycled to correlate to a can of Orange paint.
Encryption doesn't work if the content is known. To figure out the coorelation, you buy the 36C bra and scan it yourself. Encryption is not as safe as everyone belives. Randomization between the content and the signal is a reality of practice today, and it is the safest way to go, in terms of privacy concerns.
1. David Locke on June 2, 2006 7:54 PM writes...
RFID accells at closing the gap between the real and the digital. The privacy concerns stem from the notion that the signal can be correlated to the real.
The signal itself can be encrypted, but the encrypted signal can still be correlated to the real. Instead of a signal that says a 36C bra, the signal might say 342331, which always correlates to a 36C bra.
But, the reality is that the bra is acutally given a random ID. The random ID is correlated to a particular SKU only in the SKU table. This means that 342331 would coorelate to the 36C bra this week, and after it is sold, the ID can be recycled to correlate to a can of Orange paint.
Encryption doesn't work if the content is known. To figure out the coorelation, you buy the 36C bra and scan it yourself. Encryption is not as safe as everyone belives. Randomization between the content and the signal is a reality of practice today, and it is the safest way to go, in terms of privacy concerns.
Permalink to Comment