advantage frantically recruits Alfred's item

situations by allies.

Under development is even more sophisticated "topic recognition" which
can home in on guarded conversations that avoid potential trigger words.

Nothing and no one is exempt.

For example, you are talking on the telephone to a friend discussing
your son's school play. "Boy," you say sadly, "Bobby really bombed last
night," or perhaps you use the word "assassination" or "sabotage" or any
one of the key words the computer has been told to flag.

A hard copy of your conversation is produced, passed to the appropriate
section (in this case terrorism), and probably ends up in the garbage.

But perhaps the conversation is not so clear-cut or the analyst has poor
judgement. Then your name is permanently filed under "possible terrorist".
Weeks or even years later, you have a similar conversation and use the
same words; the computer filters it out again. Since this is your second
time, your name moves from the "possible" to the "probable" file.

Sound absurd? Not at all; it actually happened while I was at CSE.
[snip]

SIGINT specialists are honing their skills at monitoring digital
information. SIGINT agencies everywhere are increasingly throwing
their surveillance web over the Internet and other data networks
of interest.
[snip]


Mini-recap:

o The countries sharing intelligence: US, UK, Canada, Australia,
New Zealand

o Massive do


.

Relevant Pages
Quantcast