AI deep learning systems are not only learning to paint pictures and write blog posts.

AI is being trained on vast data sets, in order to match likely variables from countless incidents of crimes, in order to identify possible future crimes.

It’s called AI “crime forecasting,” and if it sounds like something out of the Tom Cruise movie Minority Report, there’s a reason.

Only it’s not science fiction.

As outlined by Eric Fritzvold, a Professor of Law Enforcement and Public Safety Leadership at the University of San Diego, an explosion of data gathering, together with AI machine learning and deep learning advances, have made it possible to predict the who, when and where of crimes that haven’t happened yet.

Fritzvold has noted:

“Utilizing so-called “deep learning” algorithms, programmers can train computers to analyze data from a vast array of sources and categories to actually predict when and where crimes are likely to occur. This allows agencies to properly allocate resources and increases the likelihood that officers will be in the right place at the right time.”

That vast array of data sources can include surveillance cameras, internet tracking data, digital info gathered by and from businesses and social media, medical and government gathered info, data from personal devices like smartphones and all the apps on them, and data that comes from IoT devices people wear and have in their homes, cars and workplaces, etc. 

A July 2022 study published in the journal Nature Human Behavior reported that an AI algorithm fed data on Chicago geographic areas gathered between 2014 and 2016 was able to predict crimes with 90 percent accuracy.

It was also tested with seven other cities, with similar results.

Another AI related technology that may be rolled out sooner or later to serve as a surveillance modeling and prediction: Digital Twinning.

The Chicago AI study involved digital twinning of the city, according to Engineering & Technology. But any physical environment, including objects–and-citizens–can be digitally twinned.

A digital twin is a virtual replica of a physical object, which is assigned and can continue to collect data that mirrors the movement, and moment to moment data points that the physical object reports, via different types of tracking and surveillance.

IARPA Announces Program Using AI For Intel Tips

As regional law enforcement agencies have been testing the capabilities of AI deep learning to predict crimes, Federal intelligence agencies are also funding major new initiatives.

On 25 January, IARPA (Intelligence Advanced Research Projects Activity) detailed the use of AI to help with Intelligence assessments.

Called REASON, (Rapid Explanation, Analysis and Sourcing ONline) the initiative is developing “novel technologies to help intelligence analysts substantially improve evidence and reasoning in draft analytic reports.”

According to IARPA, which operates under the Director of National Intelligence (DNI):

“REASON will assist and enhance analysts’ work by pointing them to previously unconsidered key pieces of evidence and by helping them determine which alternative explanations have the strongest support. To achieve this, REASON will exploit recent advances in artificial intelligence (AI) so evidence is provided automatically and on demand as the analyst works on a report. REASON won’t perform the analysis or write the report for the analyst, but it will seek to strengthen the report’s conclusions.”

Though the program may not be geared to outing individual pre-crimes, it’s clear that intelligence agencies are looking to integrate deep learning AI for data-driven predictive analysis.

The Trends Journal has reported extensively on surveillance technologies that are endangering the rights of citizens, including Constitutional privacy, free speech and political rights of Americans.

Some related articles include:

Skip to content