'Minority Report' moves closer to reality
- By Karen Epper Hoffman
- May 08, 2017
In the wake of terrorist attacks like those in Paris and Orlando, Fla., domestic and international law enforcement agencies alike are investigating how attackers were able to slip through intelligence-gathering networks and what can be done to prevent future attacks.
Big data analytics, machine learning and artificial intelligence technologies offer federal, state and local law enforcement agencies the opportunity to predict the probability of terror attacks based on many factors, according to David Rubal, DLT Solutions' chief technologist of data and analytics and principal data scientist. A data futurist as well as a fellow at the Institute for Critical Infrastructure Technology (ICIT), Rubal said “personal, behavioral, facial recognition, geo-location, social media and financial data” can help government agencies, law enforcement groups and their technology providers make predictions.
“Probability and risk is determined based on the intersections of this data and patterns over time,” Rubal explained.
“Agencies are also using virtual reality, derived from advanced user experiences and gaming, to simulate ‘life-like’ situations for law enforcement to improve predictability and situational awareness when training officers for responding to a terrorist threat,” Rubal said.
John Bambenek, threat research manager for Fidelis Cybersecurity, said these technologies are also used for training simulations to better prepare their people. “It’s one thing to train border-crossing agents to spot suspicious individuals by looking for undue nervousness,” Bambenek said. “It’s one step better to have them in a full [virtual training] simulation and actually looking at individuals to spot them.”
While Bambenek admitted these nascent technologies still need “to be more fully developed, … in the near future it could be possible to replay an actual terror attack in a [virtual reality] simulation to train law enforcement how to spot the actual attacker in an event that actually happened.” Law enforcement groups from the Morristown, N.J., police department to the Department of Homeland Security and the Navy have for several years used simulations to train agents, so the technique isn’t necessarily new, but Bambeneck said the technology has improved significantly in recent years.
Increasingly, Bambenek said law enforcement groups want to apply virtual reality to prevention -- allowing them to move beyond mere table-top exercises and allow a “red team” to attempt an actual terrorist attack in a virtual environment with defenders attempting to thwart them. The FBI and the Department of Defense are already testing these applications, he said.
But how are these developments -- once the stuff of fiction -- becoming operational? Much of the innovation has to do not only with advances in virtual reality and analytics themselves, but with the vast access to volumes of both broad and focused information that is now available, experts said.
“We are seeing a revolution in the scope and size of open source data,” said Mark Testoni, president and CEO of SAP National Security Services. “Whether you are in law enforcement or an intelligence analyst or operator, you need to identify people, their relationships to others and groups, their location, where they are going, what they are doing and how their patterns of behavior change.”
What’s changed? Today, practically every transaction people make -- from sending personal messages and family pictures to filing financial, medical and tax records to government’s broad geographic, demographic and longitudinal data -- is digitized, which Testoni said “offers [us] both a challenge and an opportunity.”
“Sure, we can locate bad actors and track their digital footprints, but the morass of data is overwhelming and growing immensely,” he explained. “New analytical platforms offer the power and potential to leverage this digital ocean and augment the great work being done by these professionals.”
Emerging commercial technologies now allow for analysis on the fly and can harness the digital pattern of the life of suspects in investigations, Testoni said. This makes it possible to identify and connect suspects and organizations through their activities in the open source data stores, like social media and other public domain information.
The sheer amount of potentially relevant data, coupled with increasingly improving analytics, and huge leaps in in-memory high-performance computing, are giving rise to applications that can pair with algorithmic capacities and “allow for sentiment and link analysis in scores of languages, dialects and local conventions,” Testoni said. And by cross-referencing all this against traditional methods of collection and detection, there is growing potential “to more readily identify nefarious actors and their linkages to others.” Within a matter of years, he said, law enforcement professionals should be able to get inside of the execution window of perpetrators, so they can unravel and analyze their networks, preemptively conduct apprehensions or even mitigate attacks.
According to Kenneth Geers, a senior research scientist at Comodo and former National Security Agency analyst, “virtual reality is a truly powerful technology in the evolution of games” for training for law enforcement. “Human have short lives and simple habits,” Geers said. “The basics of ‘Minority Report’ are not only going to happen, they will happen sooner than we think.”
Karen Epper Hoffman is a freelance writer based in the Seattle area.