By Jared Chausow
By Katie Toth
By Elizabeth Flock
By Albert Samaha
By Anna Merlan
By Jon Campbell
By Jon Campbell
By Albert Samaha
MIT mathematician David R. Karger says harassing individuals is foolhardy, but so is refusing to consider sensitive demographics. "This is just making your predictive capability worse," he writes in an e-mail interview. "Much more appropriate is to use the best data you've got, but to remember that probability doesn't mean certainty."
Joiner-Congleton writes, "Fundamentally, the algorithms themselves (if created in a technically correct fashion) are not the thing to fear. Rather, as in life, the things to fear are the conclusions drawn and the subsequent actions taken. Nevertheless, drawing conclusions from data is a necessary thing in life. People must do this to survive. Imagine the havoc that would be wreaked on the roads of America if we ignored the sounding of a horn on the freeway. Horn-blowing is usually associated with a dangerous event. We ignore it at our peril."
She even conceives of developing algorithms so advanced that society might intervene, to get people liable to be recruited into cells back on track before they can be seduced by elements like Al Qaeda. "There is a possibility that with sufficient information about known terrorists we could evolve to the point where we could spot terrorists in the making," she argues. "We believe that individuals can be at risk of becoming drug addicts, or joining gangs, or having affairs, or any number of things at certain times and under certain conditions in their lives. . . . Thorough and continued algorithmic investigation of terrorist behavior is very likely to shed light on their origins, and possibly lead to proactive efforts."
But there's a truly slippery slope here. We live in a nation that for months has held at least 700 peopleand possibly hundreds moreincommunicado, with no more solid connection to terrorism than that they were born in Middle Eastern countries.
Privacy may seem like a luxury in a nation at war, but that moral concept lies at the heart of constitutionally guaranteed liberties. That's why so many people are willing to fight for it. A lawsuit filed by John Gilmore, an early employee of Sun Microsystems, aims to restore the anonymity central to the freedom to travel in America. He names Ashcroft, FBI director Robert Mueller, and security czar Tom Ridge as defendants, among other officials, along with two airlines. Gilmore wants to prevent security at airports from demanding identification from him, or subjecting him to arduous and invasive searches when he refuses to provide a photo ID. The emphasis, he says, should be on strengthening cockpits and developing "fly by wire" systems to automatically land planes under threat. But our terrorism fears extend well past airlines to water-tainting, dirty bombs, suicide bombers, conventional bombing, or even simply opening fire with an assault weapon in Grand Central Stationthe kinds of attacks that are difficult to prevent in an open society.
For now, we rely on tools like algorithms, and algorithms make mistakes. Albrecht notes that in a three-month test period, the Department of Defense investigated 345 employees after a program falsely fingered them for abusing shopping privileges. In another case, an elderly woman was repeatedly stopped and questioned in airports because her name matched that of a young man already in prison for murdera glitch that may indicate CAPPS or another algorithm is using data illegally, for basic criminal investigation and not anti-terrorism. Further, supermarket records have been seized by Drug Enforcement Agency investigators looking for purchases of small plastic baggies, often used in the drug trade, Albrecht observes.
"I am not a number!" shouted Patrick McGoohan, star of the British TV show The Prisoner, when he rejected life in an idyllic village where he was held and constantly monitored. "I am a free man." Now that this nation is at war with terror, perhaps you'll remain free as long as your "Potential Terrorist Quotient" remains low enough.