• United States



Algorithm to predict at birth if a person will be a criminal?

Jul 18, 20165 mins
Data and Information SecurityInternet of ThingsSecurity

Here are two tales about 'new' tech that could potentially predict the future. One seems much scarier than the other.

Technology is not a bad thing; it’s not inherently scary. Sometimes new technology gets misused or tainted with mission creep. Most of the time, tech actually makes our lives easier and better. Here are two tales about “new” tech that could potentially predict the future. One seems scarier than the other.

Algorithm to predict at birth if a person will be a criminal

Algorithms control aspects of your life whether you are aware of it or not. They are used to come up with risk scores and even predict the future. But how would you feel about an algorithm that seems to be ripped straight from Minority Report? It would identify criminals far before they could commit a crime, since it would “predict at the time of someone’s birth how likely she is to commit a crime by the time she turns 18.”

Richard Berk, a statistician and University of Pennsylvania professor, is working on such an algorithm, according to Bloomberg. He believes it can “predict at the moment of birth whether people will commit a crime by their 18th birthday, based on factors such as environment and the history of a new child’s parents.”

There are algorithms used “to decide which blocks police officers should patrol, where to put inmates in prison, and who to let out on parole.” These are suggested to be fairer than when the decision is left to individuals with prejudices or judges having a bad day. People who support such tools claim they will do away with prejudices, but critics say the prejudices are built into the algorithms.

Berk said, “If you want me to do a totally race-neutral forecast, you’ve got to tell me what variables you’re going to allow me to use, and nobody can, because everything is confounded with race and gender.”

Bloomberg reported that Berk’s algorithms “have been used by prisons to determine which inmates to place in restrictive settings; parole departments to choose how closely to supervise people being released from prison; and police officers to predict whether people arrested for domestic violence will re-offend.” He’s even come up with a tool for OSHA to determine “which workplaces were likely to commit safety violations.” Pennsylvania will start a pilot program in the fall that uses Berk’s algorithm to decide how long to sentence a person to prison.

Since Berk’s system will be used for sentencing decisions, it was troubling to read that 29 to 39 percent of risk score predictions are wrong. Berk said, “Focusing on accuracy misses the point.” Um, a person who was labeled a future criminal would likely disagree, but a future criminal who was decided to be a low risk would be cool with it.

When talking about crime predictions, he said, “The policy position that is taken is that it’s much more dangerous to release Darth Vader than it is to incarcerate Luke Skywalker.”

For right now, Berk believes the only limit to a predict-at-birth-if-person-will-commit-a-crime tool is the data he can to feed into the algorithm.

I don’t know about you, but to me the whole concept seems pretty scary.

Chicago’s Array of Things to help predict the future

Meanwhile in Chicago, sensor-based nodes for the Array of Things (AoT) project are being installed this summer; 12 prototypes were installed previously. Here is a map of the first 42 sensors, which were dubbed a “second set of prototypes.” The sensors will supposedly help predict the future.

AoT summed it up as:

What if a light pole told you to watch out for an icy patch of sidewalk ahead? What if an app told you the most populated route for a late-night walk to the El station by yourself? What if you could get weather and air quality information block by block, instead of city by city?

The sensors will monitor air quality, climate, noise, sound vibration to detect heavy traffic, traffic congestion, pedestrian traffic and more. Eventually, there will be sensors to monitor other urban factors such pollutants, precipitation, wind, flooding and standing water. Chicago wants to be a smart city, and there will be 500 nodes by the end of 2018.

AT&T will provide the network connectivity. The press release stated, “The data collected from the sensors will be transmitted in a highly secure manner over the AT&T wireless network to a central database server at the Argonne National Laboratory, where it will be processed and released openly to the public. Researchers, city services, and technology developers will be able to use the data to address energy efficiency, traffic safety, urban flooding, air pollution and many more urban challenges.”

Unlike Berk’s algorithm meant “to predict at the moment of birth whether a person will commit a crime by their 18th birthday,” the AoT sensors and algorithms that will help predict the future are not aimed at individuals. AoT is designed “to not collect any personal or private information.”

It uses the Waggle platform “for powerful and secure remote processing of measurements before transmission of data to a central server. For example, this capability allows images such as those of standing water or bike traffic at a street intersection to be quickly analyzed within the node, with only the numerical results such as water depth or number of bikes being transmitted and publicly released.”

AoT project lead Charlie Catlett, director of the Urban Center for Computation and Data, said, “We’re interested in monitoring the city’s environment and activity, not individuals, which is why we have built privacy protection into the design of the sensors and our operating policies.”

You can view the governance and privacy policies, but it is too late to comment on them. Public comments were only allowed between June 13 – 27.

Having tech warn a person about an icy sidewalk, data that will be open to everyone and free, seems much less threatening to me than having a closed algorithm determine at birth if a person is a future criminal.

ms smith

Ms. Smith (not her real name) is a freelance writer and programmer with a special and somewhat personal interest in IT privacy and security issues. She focuses on the unique challenges of maintaining privacy and security, both for individuals and enterprises. She has worked as a journalist and has also penned many technical papers and guides covering various technologies. Smith is herself a self-described privacy and security freak.