Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

This is a note, written by a law student. Not a reason to discount the analysis, per se, but worth keeping in mind if you were thinking of skipping the article and taking the conclusion at face value.

The actual analysis starts at 367. The gist of the argument, from a quick read, is that Kyllo (the case that held that the government could not use infrared cameras to look into peoples' homes) established that the government could not use extrasensory searches to get information about what's inside areas protected by the 4th amendment. To the extent that biological vitals and medical information are protected (an uncontroversial point), it follows that using extrasensory technology to obtain that information violates Kyllo.



To expand on your summary, first the author provides one on the last 2 pages. Also a significant part of this is that these administrative searches (at airports) must be to prevent terrorism not for "ordinary" crime. FAST would make no distinction (and it's hard for it to make a distinction between any crime and mal-intentions). Further, searches of medical data must meet a higher standard and warrants are required (at least in the case of DUI blood draws, with a few exceptions).


  The system has an eighty-one percent classification accuracy
  in a laboratory test setting
That sounds pretty good -- "It catches 81% of the bad guys".

It sounds good until you realize that bad guys are an infinitesimal portion of the population.

If bad guys are 1 in 1,000 or 1 in 100,000, virtually all of the people implicated by the system will be innocent.


Specifically, since the numbers are pretty unbelieveable if you haven't seen such arguments before: if the bad guys are as numerous as 1 in 1000, then the number of false positives will be 999*19% > 189 in 1000. That is, if you are identified as a bad guy, then the odds are worse than 1 in 190 (just over 0.5%) that you are one.


Haven't read the post, but a precision on Kyllo: extrasensory searches are fine if we're talking about doing it with commonly available tools. Binoculars are fine, helicopter-mounted infrared cameras aren't. The idea being that people _expect_ the public to have certain things but not others.

Now I imagine that most ways to get vitals would be pretty intrusive and require non-common (as in, not in use by everyday people) hardware, but generally speaking Kyllo only invalidates a certain class of instrument




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: