Pearl River      The movies and television have created myths about the critical importance in crime fighting of profilers.  Jodie Foster helped set the pattern as a rookie criminal phycological profiler in “Silence of the Lambs”, and many have trod that now well-worn path since then.  For my money, the scariest outgrowth of this cop fad in the movies was “Minority Report”, which was set in the future with Tom Cruise as part of a police crew that depended on some captive women savants who engaged in “predictive policing”, where they tried to arrest perpetrators before they committed the crime.

Talking to Rachel Corbett about her book, The Monsters We Make: Murder, Obsession, and the Rise of Criminal Profiling on Wade’s World made me think that we might be a lot closer to predictive policing than we realize, even if the Cruise version is way too sci-fi.  Corbett is normally a features writer on a range of subjects for New York magazine, but she comes to this story in a too-close-for-comfort way from happy childhood memories of a guy who spent time in their house who surprisingly to her and her family killed several people.  The question that inevitably lingered for her was whether they should have been able to guess it would turn out this way or, more tellingly, whether profiling is bunko and in truth, no one would have been able to see it coming?

In short, where Corbett ends up, digging into one example after another, finds her not coming to praise profiling, but to bury it.  Her examination of Sherlock Holmes and the character’s creator Arthur Conan Doyle, books I devoured as a youth, fascinated by his powers of “deductive reasoning”, led to Doyle in real life being asked to help develop a profile for Jack the Ripper, the criminal of the day in England at the time.  This was also a classic bust.  In one example after another, profiles are less psychologically science-based than embedded with confirmation bias from the searchers.  Jodi Foster notwithstanding, the FBI ended up renaming its initial profiling unit as not behavioral science when it turned out more voodoo and fake science than anything else.  Most of the big-time serial killers were nabbed not based on the accuracy of the police profile but by random, unrelated crimes where they stumbled.

The Unabomber gets a lot of attention.  As a shy student at Harvard enrolled in a series of behavioral experiments run by a controversial professor there, now discredited, he was notorious in his own way for pushing back at the experiment’s principles.  It was hard to read about the trials without understand the potential damage it inflicted on him as a young man.  It’s well-known that it was not profiling, but his own brother’s realization that some of his printed messages were familiar from their letters and conversations, prompting him to help identify his kin.

Those failures of profiling are legendary, but Corbett finds less known and more horrible examples of the abuse that this kind of fake science can bring in Pasco County, Florida.  There the police department did try to pretzel a couple of trends, especially with young people, in significate abuses emblematic of predictive policing.  They made life a nightmare for a family when they targeted the teenage son, based on behavior that was pretty much standard teen stuff, like missing class and minor this and that.  The family had to leave the area.  The youngster was harassed in and out of juvenile detention so couldn’t finish school with his peers and instead got a GED.  Pasco paid a price, but the family paid with their lives, all of it based on the pseudo-science that criminality could be predicted and prevented based on minor behavioral psych clues and misnamed “intelligence led policing.”

Corbett doesn’t spend much time on artificial intelligence, but if she was writing the book now, I’d bet money there would be a chapter on how wrong AI could brand people.  I would also bet money that there will be police departments that try to use AI to our peril and the communities where they work.