Improving Law Enforcement Intelligence Gathering and Use with Open Source Intelligence (OSINT) and Artificial Intelligence (AI)

Natalie Faulk
Aug 24, 2018 · 8 min read

Gone are the days of the flatfoot cop pounding the pavement looking for evidence, interviewing witnesses, and spending hours pouring over reports in order to solve a crime.

As society has evolved, technology has as well, and there is a growing awareness that already-established police techniques — if used exclusively — are somewhat out-of-date and oftentimes quite expensive for what they offer. When departments sink valuable resources into maintaining old systems instead of investing into newer, more efficient, and cost-effective technologies — especially in an era of budget cuts where law enforcement agencies are forced to make difficult decisions as to where to cut funding — these agencies are missing out on a valuable source of information.

One only needs to look at history to witness the evolution of criminal investigations. Fingerprinting, DNA analysis, and computer information systems such as CODIS (Combined DNA Index System) and NDIS (National DNA Index System) have improved investigatory efforts considerably; however, as technology continues to evolve — and criminals are openly taking advantage of this new technology — law enforcement agencies may be missing out on a valuable opportunity if they don’t embrace more openly the tremendous benefits such new technology brings.

Crime is costly. The United States spends more than $100 billion annually on law enforcement and incarceration, and this figure does not even consider other economic impacts of crime in terms of victims’ costs, property devaluation, and higher outlays for companies to ensure their security.

In order to be effective, law enforcement agencies must have access to various types of intelligence. With the advent of more sophisticated technologies such as artificial intelligence (AI) and the plethora of open source intelligence (OSINT) — both of which are plentiful and inexpensive — it becomes increasingly necessary for police agencies to add newer AI and OSINT technology to their repertoire. By doing so, these agencies can enter a brave new world of more effective investigations and enhanced crime control.

Given the spate of science fiction movies and television programs, AI is a salient topic that appears to be increasing in both popularity and utility. AI is defined as the ability of a machine to learn and apply what it has learned to specific tasks. Thus, companies and various jurisdictions across the globe are experimenting with AI-assisted methods to prevent and reduce crime and to respond more quickly to crimes in progress.

AI crime prevention technologies can be divided into two broad categories: detecting crime and preventing future crime. Several companies have introduced AI technologies that have demonstrated some success in achieving these goals.

“ShotSpotter uses acoustic sensors and cameras placed around a particular location”

Using “smart city infrastructure to triangulate the location of a gunshot,” ShotSpotter uses acoustic sensors and cameras placed around a particular location that can locate in real time from where gunfire is occurring within and quickly alert authorities in a timely manner. Officers can log into a computer interface that shows the officer relevant information during an incident. This technology has the ability to improve police response to such incidents — especially in cases where 911 is not notified — toward the goal of apprehending suspects and preventing additional injuries. Of the 90 police departments that have implemented this technology, 71 percent rated its overall value as “very high.”

“Deep neural networks [that] simulate human brain activity”

Manufactured in China, Hikvision produces advanced CCTV and video surveillance equipment that utilizes deep neural networks to simulate human brain activity, especially pattern recognition. Thus, these cameras are able to use facial recognition to search for missing persons or criminal suspects, scan license plates, and detect suspicious anomalies. Of primary importance is that this type of AI technology decentralizes these functions and reduces bandwidth usage because these functions are performed in the cameras themselves and not in the cloud or some other centralized hub. Hikvision has been credited with a 65 percent reduction in crime in Sea Point, South Africa and an 80 percent detection rate of “threatening visuals” in London, England.

Another Chinese company, Cloud Walk’s AI technology seeks to predict if a person will commit a crime before s/he does so. Through facial recognition and gait analysis, the software looks for suspicious behavioral or movement changes over time and alerts law enforcement agencies.

Another innovator, Predpol’s AI technology attempts to forecast when and where crime will occur. Like more traditional crime mapping activities where high-crime areas are analyzed for patterns, this technology takes such efforts a step further. By analyzing historical crime data and current trends, it seeks to predict when and where future crimes are more likely to occur. After Tacoma, Washington, implemented Predpol in January, 2013, it saw a 22 percent decrease in residential burglaries within two years — far surpassing the city’s goal of decreasing burglaries by 7.5 percent.

Super-sleuth AI technology Visual Analytics for Sense-Making in Criminal Intelligence Analysis (Valcri) scans data to suggest the how and why a particular crime may have been committed, thus enabling investigators to devote more time and resources to building a case. Originally developed at the University of Middlesex in London, England, as an early warning system for impending criminal activity, Valcri has been expanded to take on more routine police work. By quickly analyzing reports, images, and videos, Valcri looks for links that warrant further investigation in a much shorter timeframe than humans are capable. Future plans for real-time notification are in the works.

Known for a variety of AI and other tech services, Intel (the chipset maker) has created AI technology specifically geared toward locate missing children. In conjunction with the National Center for Missing and Exploited Children (NCMEC), Intel is modernizing how NCMEC organizes and examines data, investigates tips, and reports to other law enforcement agencies.

“If criminals can be identified before they offend, the subsequent costs associated with crime are also diminished substantially.”

While AI technology does come with a price tag, the significant reduction in man hours more than offsets the initial expenditures. Additionally, if criminals can be identified before they offend, if victims can be reduced, and if criminal hotspots are identified and neutralized, the subsequent costs associated with crime are also diminished substantially.

As is the case with many technologies and advanced crime-fighting tools, the question of privacy is prevalent. Given the dramatic increase in security cameras in recent years and the sheer amount of data being disseminated to law enforcement and security agencies, this is a salient concern. Civil rights agencies are always on the offensive regarding presumed interference into constitutional privacy rights; however, there will always be a necessary tradeoff between security and privacy. It becomes critical to balance these concerns in a way that is beneficial to all involved, and crime reduction is a necessary endeavor.

“OSINT can assist with categorizing and verifying facts, identifying criminals, understanding ideologies”

Open source intelligence (OSINT) is any intelligence derived from publicly available information sources that anyone can “lawfully obtain by request, purchase, or observation.” From a law enforcement perspective, OSINT can assist with categorizing and verifying facts, identifying criminals, understanding ideologies, and recognizing emerging crime trends. In recent years, the availability of OSINT has exploded. In fact, this type of intelligence provides a significant amount of information used by intelligence and security agencies to identify risks and make effective and timely strategic decisions.

There are several pros and cons of using OSINT. Perhaps the biggest advantage OSINT is the fact that it is far less expensive than more traditional information-collecting tools, thus offering a greater potential return on investment (ROI). Given that when states, counties, and municipalities are faced with budget cuts, law enforcement resources are oftentimes on the chopping block, it makes sense to embrace technologies that are not only effective but also less expensive.

Another benefit of OSINT is that information can be easily — and legally — shared with other agencies. The best sources of OSINT include periodicals, newspapers, television and radio broadcasts, public databases, directories, open discussions and forums, government reports, newsletters, blogs, social media platforms, commercial information from vendors, search engines, RSS feeds, and websites. In fact, the rapid growth of the Internet has made it a major source of knowledge. Of particular concern, however, is to understand that quantity does not necessarily mean quality, so any open source information should be viewed with a critical eye for credibility and veracity.

A wealth of knowledge can be found in OSINT. Extremist individuals and groups oftentimes want to share their goals and beliefs, so they tend to post frequently in a variety of media to spread their message. Further, criminals enjoy the anonymity they believe the Internet provides. Social media platforms have become a central hub of criminal indoctrination and terrorist radicalization, and investigators would be foolish to not take advantage of this incredible source of intelligence.

Further, because open sources are referred to as sources of first resort, this type of intelligence poses a minimal intrusion on civil liberties. Especially with respect to social media platforms and other public online forums, that individuals and groups post willingly to these sites essentially negates any future claim of civil rights infringement.

“Even if the initial information may not be the best, it can provide a starting point for additional investigation.”

One of the biggest shortcomings of OSINT is the potential for information overload due to the tremendous amount of “noise” on the Internet. Thus, locating the correct information can become quite time-consuming. Additionally, OSINT is not generally ready-to-use. Instead, analysis of raw data is necessary to identify credible, valid, and verifiable information. Even if the initial information may not be the best, it can provide a starting point for additional investigation.

Another obstacle involves the potential for analysts to lack adequate subject matter expertise in identifying appropriate OSINT. Relatedly, one’s inherent bias against such information may also serve to reduce the potential of such information. Other potential barriers include the lack of effective analytic tools and the fact that multiple media outlets report on the dame story, thus potentially skewing a particular source and giving it more credibility than warranted.

Despite these potential shortcomings, OSINT is a critical tool that should be in every law enforcement agency’s repertoire because of the potential wealth of valuable information therein. Over the years, intelligence agencies have been criticized for failing to take advantage of available OSINT. From the 1996 Aspin-Brown Commission to the 2001 9/11 Commission to the 2005 WMD Commission, increased use of available OSINT could have led to dramatically different outcomes.

“why police agencies have been hesitant to take the plunge to make AI and OSINT a growing part of their current activities?”

Given the evolution of criminals and the spread of deviant ideologies, technology has had to keep pace. Thus, it makes sense for law enforcement agencies to stay abreast of new methods to obtain intelligence in order to enhance their identification and apprehension of criminals and prevention of crime. That the military and security agencies have openly embraced OSINT and AI as both force and resource multipliers to improve their intelligence collection efforts begs the question as to why police agencies have been hesitant to take the plunge to make AI and OSINT a growing part of their current activities. While not a complete substitute for more traditional intelligence gathering efforts and policing activities, OSINT and AI provide significant cost-effective enhancements that free up valuable resources in terms of time and manpower to improve an agency’s efficacy.