Wednesday, 22 October 2003
Terrified of Terror Profiling?
Bruce Schneier, the renowned expert in computer security (as well as founder and Chief Technical Officer of Counterpane Internet Security, Inc.) wrote a column this week for Newsday: "Terror Profiles By Computers Are Ineffective." As the title suggests, Schneier argues that all the approaches yet taken to "profiling" terrorists suffer from the same fundamental design flaw. "There's a common belief — generally mistaken — that if we only had enough data we could pick terrorists out of crowds," Schneier writes. He goes on to show that the types of information that we have endeavored to gather — indeed, the types of information that we can gather — bear no statistically significant relationship with terrorist acts, or even propensity toward terrorism.
Schneier's argument is bolstered by the simple, elegant, and compelling mathematical analysis done by Temple University mathematician John Allen Paulos, in the January 2003 installment of his column "Who's Counting?." The article, "Future World: Privacy, Terrorists, and Science Fiction," assumes that a project such as the recently de-funded Terrorist Information Awareness program (née "Total Information Awareness"), has succeeded beyond the wildest dreams of its founders by 2054, the year when the film Minority Report is set. This hypothetical program has a predictive success rate of 99%. Examining this number and assuming that the U.S. has 300 million citizens, Paulos proves that it would imprison just under 1,000 terrorists and just under 3 million innocent people.