I'm in NYC, attending the 2003 version of the
Computers, Freedom, and Privacy
conference. No live blogging of it for me this year, but here's a reaction or
so to some events yesterday.
Listening to the presentation in defense of the Total Information Awareness
program (TIA), I had two thoughts, or inights, or gut reactions:
First, I was particularly struck by the analogy
drawn with high tech weapons of war. (You can download an mp3 file of
Total
Information Awareness - A Debate.) We were told that TIA was great
because, like smart bombs, it would allow law enforcement resources to bore
right in on the evildoers. Opponents were labelled "Luddites" for opposing it
just because there would be false positives, since false positives are
inevitable.
Of course, we know there are false positives with smart weaponry, as well.
Bit by bit, we're learning to call them "collateral damage."
So, I found myself wondering, how good is TIA? Assuming, for the sake of
argument, that it does exactly what it is advertised to do (a long, long,
long shot), would the proponents please quantify its great value in terms of
the collateral damage justified. What burdens are innocent Americans
supposed to bear?
Second, I realized that the proponents of Total Information Awareness
are really offering a version
of Pascal's Wager. Blaise Pascal argued that one ought to try to believe in
the Christian God, since there was a nonzero probability that the God
exists, infinite benefits for believing, infinite costs to disbelief, and
only finite costs for belief and finite benefits (if any) to disbelief. THe
argument here seems like this: The benefits of TIA are (approximately?)
infinite, ergo they justify *any finite cost* no matter how large.
But - the benefits are not really infinite, and
- the argument is open--
as
is Pascal's--to being countered by a 'Many Gods' argument.
Pascal's Wager
argument, if it works for the Christian God, works for the existence of any
God that promises infinite rewards for belief and infinite punishments for
disbelief. We can't possibly believe or try to believe in the indefinitely
many such Gods. Similarly, indefnitely many projects, of which TIA is but
one, offer a nonzero probability of (approximately) infinite benefit in the
War on Terrorism and so on. But we can't undertake them all. So it follows
that we still would require some reason and analysis to justify or motivate
our belief in this program.
That's my little riff on the TIA debate.
(Oh: the opponents of TIA -- Katie Corrigan and Barbara Simons -- roundly
defeated to proponents -- Heather MacDonald and Michael Scardaville. Best
question from the floor was Patrick Ball's, quickly doing the math to show
that, given the small number of cases of terrorism from which to build the
patterns to be used in TIA's data mining the number of false positives has to
be just huge.)
8:42:43 AM 
|
|
