8 Companies Using AI for Law Enforcement
Table of contents
The science fiction universe is populated by all sorts of crime-fighting artificial intelligence (AI) machines. David Hasselhoff had KITT (Knight Industries Two Thousand) in Knight Rider. Tony Stark, aka Iron Man, relies on JARVIS (Just A Rather Very Intelligent System). RoboCop was a well-armed fighting cyborg that melded man and machine. Today, researchers and companies are turning fiction into reality by building AI to fight crime.
In Europe, for example, scientists are developing a system called VALCRI (Visual Analytics for Sense-Making in Criminal Intelligence Analysis), a name alone that might strike fear into the hearts of criminals, even though we have no idea how you get an acronym like VALCRI with that convoluted name. VALCRI is designed to do all of the data crunching for human detectives, finding clues faster even than Scooby Doo (that rhymes!).
How? VALCRI can scan millions of police records, interviews, pictures, videos and other types of evidence to identify patterns and make connections that might be relevant to an investigation, according to a feature in New Scientist. A crime analyst on a new case must gather information from dozens of databases to find connections between incidents based on things like location or modus operandi. VALCRI can do the same job with a single click, New Scientist reported. It can understand subtleties of human language and use facial recognition software to search for bad guys in video and pictures. A different project, out of Scotland, trained algorithms on episodes of the TV crime series CSI. After assimilating images, dialogue and scene descriptions, the machine identified the killer 60 percent of the time versus 85 percent for humans.
Meanwhile, companies and startups are already starting to commercialize AI technology that will help the McGruff Crime Dogs of the world take a bite out of crime. Below we highlight eight companies to demonstrate how law enforcement might look in the not-too-distant future.
VALCRI for Realz
One company that is developing something that sounds very similar to VALCRI will be familiar to regular readers—Veritone (NASDAQ:VERI). We became very interested in the Newport Beach, California-based company when it went public this year, offering one of the few opportunities to invest in what, on the surface, seems like a pure play in AI. Our spidey sense did go off recently when we saw the roller coast ride the company’s stock has had on the NASDAQ since it went public, making it look like a possible pump-and-dump stock. A quick recap of what they do: Veritone’s AI platform integrates dozens of different “cognitive engines” from companies like Google and IBM, a well as its own in-house algorithms. The idea is that customers send their data to Veritone for processing through its cognitive engines, which include services like transcription, face detection, object detection, sentiment analysis, geospatial data, and audio/visual “fingerprinting” to name a few.
All of that sounds like a good combo for law enforcement, and that’s how Veritone is marketing its AI platform to governments. The company says it can extract information such as faces and license plates, tagging each piece of evidence with a location, date and time. The system also works like Sherlock Holmes, finding patterns that would elude the average Lestrade. And, as is typical with machine learning, performance improves the more it is used.
Company officials at Axon are apparently big sci-fi fans. The company describes its Seattle headquarters as a “mix of Star Wars, James Bond, Get Smart and Star Trek”. Oh, never heard of Axon? Maybe you’re more familiar with the company’s former name—Taser International. The company is apparently pivoting from manufacturing devices designed to deliver free heart attacks to a Minority Report-style service to law enforcement.
A website called The Intercept has written an excellent teardown of Axon’s plans for world domination, but we’ll give you the thumbnail version here: Aside from tasers, Axon also manufactures body cameras for police officers, and earlier this year it made an offer they couldn’t refuse. Free body cams for every police officer in America. Their motivation: Data, untold terabytes of data. Axon made a couple of AI computer vision acquisitions earlier this year. Imagine being able to identify a wanted suspect in real-time as video streams into the servers at Axon. In a longer, dystopian view, the company can train algorithms to forecast who might be a criminal based on looks alone.
CSI for Computers
Not every criminal case involves physical forensic evidence like blood stains or fingerprints. In this day and age, many investigations rely just as heavily on digital forensics. We’re not just talking about uncovering porn searches on a cleared web browser but things like extracting information from encrypted servers. A 30-year-old company called AccessData, which raised $45 million in 2013, has developed a platform for managing digital forensics investigations called AD Lab. AccessData does not specifically call out AI or machine learning on its website. However, an analysis of the company’s system notes in an article titled, “The Coming AI Revolution in Digital Forensics,” that AD Lab helps investigators “conduct searches for documents, images, addresses, social security numbers, etc., and the tool then guides the examiner through the analysis and processing of the data”.
A more explicit example of machine learning in digital forensics can be found in Waterloo, Canada-based Magnet Forensics. The startup’s flagship product is Internet Evidence Finder (IEF), a software platform that helps investigators find and analyze digital evidence across multiple computer and mobile devices. Its newest offering, however, is Magnet.AI, a contextual tool that uses machine learning to sort through conversations on smartphones, computers and chat apps, according to a story on Startup Toronto. The tool is designed specifically to assist investigators working on child exploitation cases.
In a similar vein, Pittsburgh-based Marinus Analytics uses AI technology to help law enforcement track down missing persons, particularly those involved in sex trafficking. Founded in 2014 and spun out of Carnegie Mellon University Robotics Institute, Marinus originally deployed a machine-learning tool called Traffic Jam to analyze patterns from data such as cell phone numbers listed in online sex ads to track down criminals involved in human trafficking. Law enforcement agencies including the FBI and major municipal police departments use the software. More recently, Marinus deployed a new tool called FaceSearch, powered by Amazon Rekognition, which allows law enforcement to search millions of online records to find victims using facial recognition.
The Real Thing
Founded in 2012, New York-based Entrupy has raised $2.6 million for its on-demand system that verifies the authenticity of high-end goods like Gucci and Prada. Another university spinoff, Entrupy has developed algorithms that allow it to analyze various materials ranging from canvas and leather to metal and wood. The device, which looks like a handheld scanner, takes microscopic photographs of different areas of an item and runs them through a computer. Entrupy is accurate more than 98 percent of the time and returns results in less than 30 seconds. Customers receive a certificate of authenticity, like the one below, if the item is the real deal.
A little more about the bag featured above. Until recently, it had been an assumed fake by the store, forgotten under a pile of merchandise for two years. After the client started using Entrupy, it decided to test the “fake” Chanel bag. Turns it was authentic and worth $2,000. Two grand for a purse? That’s the real crime here.
There’s an App for That
Founded in 2015, Staqu Technologies out of India applies AI techniques such as deep learning and computer vision to fight both crime and fashion faux pas. There are quite a few numbers floating around out there about how much money they’ve raised but we’ve learned from the horse’s mouth that they raised $500,000 in seed funding. The company has developed an app for police officers called ABHED (Artificial Intelligence-Based Human Efface Detection) that reportedly allows them to search criminal registries, conduct missing person searches and even match fingerprints at a crime scene. The company claims to have helped solve 30 cases in one police district alone.
It also leverages its AI technology to help consumers find that perfect evening gown or suit by quickly searching through millions of images online, returning better matches with each use. Yet another way AI is changing fashion sense.
Caught on Tape
Founded in 2016, San Francisco-based Deep Science AI has developed what it calls an AI surveillance (AIS) platform for businesses. AIS uses deep learning to identify in real-time people concealing their faces, firearms, or intruders after-hours or where they shouldn’t be, and alerts a security analyst monitoring remotely. The platform allows the analyst to monitor 500 or more surveillance feeds at once, according to the company. The service comes as cheap as $44.99 per camera for an annual contract.
It doesn’t take an AI Sherlock Holmes to see a pattern here among many of these law enforcement solutions. Companies are leveraging advances in deep learning, particularly in computer vision, to identify objects like guns and people with criminal records with unprecedented granularity. On one level, this will allow the good guys to catch the bad guys in the act. But, being human, we’ll want to take it further, using these tools to assess emotion and intent, if not outright stereotyping to identify someone as a criminal based on his or her looks. It’s unsettling to think that companies like Axon are already moving in that direction but terrifying to consider an entire country like China is already investing in thought crime prevention using artificial intelligence.
Become a premium member and get access to hundreds of premium articles, reports and additional content.
Nanalyze Premium is your comprehensive guide to investing in disruptive technologies. Read by the top investment banks, management consultancies, VCs, and research houses. Trusted by over 100,000 institutional and retail investors. Covering disruptive technologies for nearly two decades.