2 ms·
My concern would be that if the wrong disks get imaged (either by real bumbling, or erring on the side of over-collecting, or calculated-hoovering-that-could-be
by gojomo 15y ago
My concern would be that if the wrong disks get imaged (either by real bumbling, or erring on the side of over-collecting, or calculated-hoovering-that-could-be-portrayed-as-an-honest-mistake-later), the disk data could then wind up in some broader forensic analysis pipeline, or long-term evidence archive, and then be eventually misused.
For example, if they think they may have imaged too much, do they promptly and irreversibly wipe the extra? Or keep it, just in case their assessment of its relation to the current investigation changes again? Could current or future automated criminal-activity-analysis engines be run against every old disk image in their possession, ignoring the details of how/why they were collected?
- wmf 15y agoCould current or future automated criminal-activity-analysis engines be run against every old disk image in their possession... As problematic as that is, it's a drop in the bucket compared to things like Echelon. I think people should be realistic about their paranoia.
- nknight 15y agoEchelon can be guarded against with in-transit crypto. Physical access to systems is far more problematic. Even if the data is nominally encrypted on-disk, the key must be in RAM to make use of the data, rendering it vulnerable at the time of seizure. (You can partially mitigate the risk, but you can't make it 100%, and it's very likely you'll have issues with cost, performance, and user-friendliness along the way.)