The discussion of “big data” encompasses many things. As more and more of our lives are recorded and parsed, the potential for machine learning and artificial intelligence will become ever greater.
From the realm of transportation, with the development of autonomous vehicles to the arena of criminal justice, the collection and processing of ever more discrete and minute elements of everyday life holds great promise to improve our lives or create a more suffocating level of control.
Companies such as Microsoft are already marketing their wares to law enforcement, in an effort to capture what could be a very large and lucrative market, selling predictive policing systems that purport to suggest to the police those most likely to be involved in future crimes.
At the moment, these systems are relatively rudimentary, analyzing police records of arrest and service calls. It may be somewhat useful in alerting police to patters that would take human analyst hours or days to uncover. It may show a pattern drug deals or burglaries, and lead to a greater police presence to disrupt that pattern.
They are poor at more random, private crimes such as murder, and actually show more promise for alerting police to those who are most like to be victims of crime.
As data collection becomes more comprehensive and the algorithms become better, they will likely be sold to other law enforcement or agencies like the Internal Revenue Service, which could use them to sift through the millions of tax transactions they process every year.
Such a system could greatly increase the likelihood of tax audits and the potential for criminal charges to those who fit a pattern.
But these programs will need to be closely monitored. Trusting law enforcement alone to ensure these systems are used correctly is bad public policy, as any system that produces more convictions would be viewed as successful in their eyes.
The danger is that such systems could take on a life of their own, creating data to validate their own use.
Source: csmonitor.com, “Microsoft says its software can tell if you’re going back to prison,” Max Lewontin, November 6, 2015