Content provided by Cognyte
In several recent public appearances and speeches, FBI Director Christopher Wray has used the same phrase to describe the overall public safety threat level, saying that he sees “blinking lights everywhere” across the entire range of threats the FBI is monitoring – from foreign terrorism to domestic terrorism, drug traffickers, cybercriminals and spies.
It’s an apt phrase.
It’s also an accurate description of one of the most critical challenges currently facing virtually every law enforcement agency in the U.S. and other developed countries – the challenge of multiple, simultaneous threats to public safety that risk overwhelming police agencies’ resources. One of the key factors driving that mismatch between resources and threats is the volume of data law enforcement agencies have to analyze and manage.
Police and other law enforcement agencies are drowning in data – from mobile phones, laptops, security cameras, body-worn cameras, videos, social media, case files, dispatch systems, telecommunications records and many other sources.
The current generation of smartphones holds 128 to 512 gigabytes of data. It’s helpful to use a non-digital viewpoint to understand what that means. One gigabyte equates to 50,000 to 75,000 pages, which is about 100 to 150 reams of paper or roughly 10 to 15 cases of paper – enough to fill a small truck. So, 512 gigabytes equate to more than 500 small trucks filled with paper.
The only way to meet this challenge is by applying advanced analytic technology and artificial intelligence (AI).
FBI Director Wray expressed the importance of AI tools in this way in a news media interview while visiting an FBI field office: “AI is in many ways the most effective tool against the bad guys’ use of AI,” he said. “So we need to work closely with industry to try to help make sure that American AI can be used to help protect American people from AI-enabled threats coming the other way.”
There is almost complete agreement among law enforcement professionals that AI tools offer tremendous benefits. In early 2024, Cognyte conducted a research survey of 200 law enforcement professionals worldwide and 99% of respondents said they consider AI beneficial for law enforcement data analysis.
When asked what specific benefits they expect from AI-powered data analytics, the top responses included better identification of suspicious patterns and hidden connections (54%), enhanced data processing speed and efficiency (49%) and increased scalability for handling larger datasets (49%).
“Decision intelligence” is a term now used to describe the type of AI-powered, advanced analytics tools we’re talking about. Specifically, what are the capabilities of the best decision intelligence tools and how do they help law enforcement professionals speed up their investigations, improve efficiency and make unprecedented amounts of data an advantage instead of a burden? Let’s take a closer look.
Advanced search and filtering
Decision intelligence technology enables investigators to find the proverbial needle in the haystack and to do so with incredible speed. An investigator may want to start with a single data point: A phone number, a location at a specific time, a vehicle license plate, a suspect name or “street name” or a photograph. Decision intelligence technology uses entity resolution and link analysis to find any associations with that starting data point, and – depending on the volume of data being queried – it does so in seconds or minutes instead of days or weeks.
Thanks to large language models (LLM), investigators don’t have to be computer scientists to write sophisticated queries. They can use common words and ask simple questions, just as they would when using an online search engine like Google.
Scoring and prioritizing risks
When you’re seeing “blinking lights everywhere” and experiencing an overload of data and risks demanding attention, it can be demoralizing and overwhelming. What should my immediate priorities be? What are the most important cases and actions to address? What are the best opportunities to solve a case, disrupt an ongoing criminal enterprise or prevent a planned terror attack?
With AI-powered risk scoring, decision intelligence technology can analyze vast volumes of data, quantify every relevant factor and provide actionable insights that give clear answers to those questions. Examples include: The mass drug shipment that intelligence sources expect to enter the port next week is most likely to be on this ship. The person of interest with the closest association with the victim is X.
Leveraging LLMs enables far stronger capabilities than traditional text analytics. This is because the analysis extends beyond the text in any specific document. An LLM will leverage its extensive knowledge base, based on the vast datasets it has been trained on, to create new entities, infer implications and establish connections that may not be explicitly stated in the text of that specific document. For example, a police report may have a minor reference to a specific gang with no additional details, but an LLM could tap into all the data on that gang available on the web.
Easy integration of new data sources
Investigations, cyber security defense and intelligence work are dynamic. Things change rapidly and analytic insights lead in new directions. New data sources become available and need to be analyzed. Typically, analysts must spend significant portions of their time preparing and processing data for analysis before they can query it and gain insights. Decision intelligence solutions can dramatically accelerate those processes, making fusing and enriching large volumes of data of diverse types and sources easier and faster.
Automated workflows
Decision intelligence can also automate time-consuming analytic processes to improve efficiency and accelerate the pace of investigations. For example, AI can automatically comb through thousands of hours of video and audio footage, flagging relevant objects like weapons, drugs and vehicles. It can also use facial and voice-matching capabilities to identify persons of interest and associations between them and events. Automated content analysis can also leverage Optical Character Recognition (OCR) to analyze textual elements in videos, such as signs, license plate numbers and text on documents.
A real-world example and success case
Hundreds of government agencies use decision intelligence technology worldwide and have experienced improved outcomes. Most cases have to remain confidential because of the nature of their work. Still, details can be revealed in a few cases, such as the following: A government agency’s financial intelligence unit (FIU) was responsible for investigating and prosecuting financial crimes and money laundering. Investigators there had access to large volumes of diverse data sources, but they had been unable to fuse information from the disparate sources into one unified view. After implementing Cognyte’s NEXYTE decision intelligence platform, they were able to bring together all their data into a single investigative workspace and generate enriched profiles for every financial entity of interest, including beneficiaries, accounts, companies, transactions and assets, and map the connections between entities.
This allowed investigators to proactively discover hidden links between beneficiaries and shell companies, as well as detect illicit financial activities. As a result, investigators were able to identify a previously unknown “straw man” who was playing a key role in a major money laundering operation.That result was achieved within three days – in a case that the organization had been investigating for three years.