CIA patching ECHELON shortcomings
Relax, it's for your own good, and the sake of your children
A core objection to paranoid rants regarding the US National Security Agency (NSA) electronic eavesdropping apparatus called ECHELON is the simple observation that spooks trying to use it are literally buried in an avalanche of white noise from which it's quite difficult to extract anything pertinent.
But now the US Central Intelligence Agency (CIA), no doubt with some assistance and guidance from NSA, is making strides towards cracking that little inconvenience.
The CIA's Office of Advanced Information Technology is developing a number of data-mining enhancements to make life easy for those who would eavesdrop on electronic communications, Reuters reports.
First up is a computer program called Oasis, which automatically converts audio signals into conveniently readable, and searchable, text.
And it distinguishes voices, cleverly enough, so that the transcript of an intercepted 'broadcast' (a conference call via mobile phones?) will show each speaker automatically identified as Male 1, Female 1; Male 2, Female 2; and so on.
If the transcript seems implausible at any point, or disappointingly mundane, the operator can easily listen to relevant parts of the broadcast to check the machine's accuracy, and determine that "recognize speech" actually was "wreck a nice beach," and send in the appropriate goon squads to prevent it.
Oasis also references search terms and keywords automatically. Thus text containing the phrase "truck bomb" would pop up in a query for "terror*".
The CIA is planning to develop Oasis for spy-useful foreign languages such as Arabic and Chinese, the wire service says.
Next comes a software tool called FLUENT, which enables an operator to search stored documents in a language s/he doesn't understand by using his or her own language for queries.
So, imagine an uneducated, highly-trained Anglophone with a PhD in some anti-intellectual pseudo-'discipline' like medicine, education, women's studies, engineering, creative writing, economics or computer science, naturally poorly acquainted with languages, but charged with grave national security responsibilities.
Say this person needs to know what the Chinese have been publishing about nuclear warheads.
Salvation: FLUENT allows them to search on "nuclear warhead" in English, and still dredge up Chinese (or whatever) documents for people with actual language skills to translate and evaluate for them. Is that cool or what?
Presently, FLUENT can translate Chinese, Korean, Portuguese, Russian, Serbo-Croatian and Ukrainian, Reuters says.
The omissions are almost as telling as the languages included. What, no Japanese, no Arabic, no Spanish, no Hebrew, no French? Laotian and Navajo we can understand, but what have we here? Laziness, discrimination, or misplaced trust?
You make the call. ®
Sponsored: Beyond the Data Frontier