Paranoid America–by which I mean its governors–has long dreamed of foolproof technology to guard the Homeland from subversion, or penetration by alien hostiles.

In its latest variant the vaunted technology comes in the form of sweeps by the computers of the National Security Agency, programmed to intercept hundreds of millions of phone, e-mail and fax messages. These days as much as a third of global communications are on fiber-optic cable routes that pass through the United States.

The NSA’s programmers claim that the artificial intelligence programs–terabytes of speech, text and image data–monitoring the filters are of such refinement that they can determine the sex, age and class of the communicators and, no doubt (though they take care not to boast of any such profiling), their genetic and linguistic ethnicity too. After all, Middle Easterners are surely a prime target.

A very useful story in the February 5 Washington Post, headlined “Surveillance Net Yields Few Subjects,” cites “knowledgeable sources” as saying that about 5,000 Americans have had their conversations recorded or e-mails read without court authority. Of these, fewer than ten US citizens or residents a year “have aroused enough suspicion during warrantless eavesdropping to justify interception of their domestic calls, as well.”

Such intercepts would require a warrant from a judge, with the request couched in terms of probable cause, usually defined as being a one-in-two chance of the suspicions being justified. So clearly a final cull of ten or so a year out of hundreds of thousands, or most likely tens of millions, means the “probable cause” standard was abandoned.

“Data mining” by artificial intelligence is a proceeding that is not only constitutionally illegal but a technological fantasy. The Post quotes Jeff Jonas, chief scientist at IBM Entity Analytics, as saying pattern-matching techniques that “look at people’s behavior to predict terrorist intent are so far from reaching the level of accuracy that’s necessary that I see them as nothing but civil liberty infringement engines.”

Every era produces its techno-Panglosses, eager to guard America and demanding torrents of public money to that end. In Reagan’s time it was the Strategic Defense Initiative, with missiles programmed to launch on warning that enemy warheads were plummeting into the Homeland. That spasm of military Keynesianism has thus far merely cost money. Back in the early part of the twentieth century the data miners and Star Wars fantasists had their equivalents in men of intellectual eminence who successfully agitated for filters to be installed at America’s ports of entry to detect genetic terrorists, i.e., people of bloodstock deemed by the fearful eugenicists to be a threat to America’s gene pool.

The US Immigration Act of 1924 sanctioned the use of the bogus US Army IQ scores of World War I (themselves promoted by eugenic racists) to “scientifically verify” the supposed hereditary mental inferiority of Jews, Italians, Poles, Hungarians, Spaniards and other non-Anglo-Saxon Protestant racial and ethnic groups. The eugenicists even wanted to send inspectors to Europe to examine all would-be immigrants for genetic flaws. In the end, this task passed to one of the eugenicists’ German admirers.

In his great 1975 tract The Legacy of Malthus, Allan Chase, narrating this shameful story, asks, How many of the 6,065,704 would-be immigrants excluded by racial quotas set by the eugenicists survived the war? For sure, most of the Jews, Poles and Russians identified by the Nazis (using US eugenic “science”) were rounded up and exterminated.

To the phrenologists and genetic data miners we can add the forensic fingerprinters. I’ve long believed that the “scientific certainty” of unique fingerprint matching is mostly theater, using suspect forensic work to bewitch judge and jury, as it has for more than 100 years. Fingerprinting, be it recalled, was first sold as a crime-fighting tool by Charles Darwin’s cousin Ernest Galton, a fervent eugenicist.

In 2004 the FBI’s top fingerprint analysts, subsequently buttressed by an outside “forensic expert,” insisted that a print lifted from a bag at the scene of the Madrid terror bombing was “a 100 percent positive identification” with one of twenty sets of prints spat out by the FBI’s Integrated Automated Fingerprint Identification System, containing a database of some 47 million fingerprints. (To be fair to the IAFIS computer, it said “close non-match.”)

The print thrown up by the FBI’s computer belonged to the left index finger of Brandon Mayfield, a lawyer working in Beaverton, Oregon. A federal judge in Portland duly acknowledged probable cause in signing a warrant for surveillance of Mayfield. He was spied upon and arrested. All the while, the Spanish police were insisting there was no match between Mayfield’s print and the one on the bag, which they determined belonged to the right middle finger of Ouhnane Daoud, an Algerian national living in Spain, whom they duly arrested. Mayfield, who was nowhere near Spain when the bombs went off, went free.

The claims of scientific precision are as suspect today as they were a century ago, when men like Charles Davenport were laboring on their racist tracts and the sterilizers were mustering strength here in America. These days we have data mining, “100 percent certain” DNA hits, retinal ID, face-recognition systems. Elementary constitutional protections get swept aside. As the Democrats reviewed the NSA data mining, a prime concern was the potential liability of US phone carriers (who poured money into their campaign treasuries in 1996 to purchase telecommunications “reform”). They didn’t question the very premises of the data mining. Is this strange? Not in a world where the New York Times can publish an article, as it did on February 8, about the Democrats’ failure to gain popular traction, in which the difficult words “war” and “Iraq” barely intruded.