NCIS Analyzes, Consolidates, and Structures Navy Criminal Investigation Data for Better Results

MoD criminal investigations have always been about gathering evidence, but traditional investigative methods are now merging new forms of data analysis. The Naval Criminal Investigative Service (NCIS) sees several changes coming as a result.

NCIS, a federal civilian law enforcement agency, supports the Department of the Navy from Sailors to Marines, placing it in a unique position to investigate crimes, prevent terrorism, or protect military secrets. Chief Information Officer Laukik Suthar said the agency has just over 2,000 people in around 190 locations around the world, so one of the main focuses at the moment is access to the Ministry of Navy for Special Agents.

Part of this access is ensuring that users have a discovery mechanism or index so that they know what data they need to work with.

“You’re right to unbox it and make sure the information is there, but we also need to move to accessibility, which means it doesn’t matter where you are, when you have access to the information, which means it could be via hardline on base, or it could be live,” Suthar said on Federal Monthly Insights – Special Bulletin: Digital Surveys. “You might be right there outside a witness’s house and you need access to that information. Or you could be at a crime scene, and you’re trying to connect the different types of information looking at you, is there a 3D rendering of the room you need? »

He said the three big components of the data challenge for NCIS include the issue of access and the issue of reliability — consolidating information so that special agents can see it based on some form of reliable information. Number three, he said, is to overlay data analysis.

“I’m not talking about artificial intelligence in its true form,” he said on Federal Drive with Tom Temin, “but linking the relevant information together, stitching it together, and then providing some kind of dashboard or context on how it’s put together for the operator in the field to make sure they have the information he needs.”

It is difficult to integrate unstructured data for a survey into a consolidated form. Suthar said the data can present itself in a very “inhuman and recognizable” way, which is unnatural for agents to watch, but makes sense for the computer.

“Machine learning and AI – artificial intelligence – are doing a really good job. Again, it’s about tooling the product well enough to read the information. But that’s something something we don’t talk about too often is not doing it alone,” he said. “You have to partner up. You can’t do it alone. You have to partner up with people and you have to work with other entities to make sure that unstructured data – structured data is exactly the same as it would be with a partner organization, because we at NCIS, we’ve worked in multiple agencies.

Mark M. Gagnon