Research on artificial intelligence technology promoting the transformation of American intelligence work: optimization of intelligence production process (1)

In recent years, with the rapid development of artificial intelligence technologies such as "deep learning" and "intensive learning", artificial intelligence technologies have developed many applications in the field of military intelligence. Modern artificial intelligence technology has completely broken through the limitations of traditional artificial intelligence, and has played a huge role in data processing, pattern recognition, robotics, gaming, data mining, machine-assisted analysis, and cyberspace security. For the U.S. intelligence community, the development of artificial intelligence technology that can be directly applied to intelligence work is the focus of improving intelligence identification, intelligence collection, data processing, intelligence analysis, intelligence sharing and collaboration.

Artificial intelligence technology promotes the optimization of intelligence production process: artificial intelligence technology improves intelligence collection and data processing capabilities.

In the context of the information age, while responding to traditional security threats, U.S. intelligence agencies also pay attention to the identification and control of non-traditional security threats. Entire societies of nations create vast amounts of structured and unstructured data at a scale that can overwhelm previous models of intelligence analysis and threat identification. The US Defense Innovation Council asserts that whoever collects and organizes the most data about themselves and their opponents will maintain a technological advantage. Failure to treat data as a strategic asset is giving away valuable time and space to competitors or adversaries.

The person in charge of integrated analysis and methods of the US Defense Intelligence Agency pointed out that there are four problems in the construction of military intelligence big data projects: poor data consistency, extensive data storage and access, low data interpretation, and incompatible data systems. The above four problems are the problems of intelligence big data, and they are also ways to realize unified data standards and give full play to the value of intelligence data.

In order to solve the above problems, the U.S. intelligence community actively uses artificial intelligence technology to promote intelligence data management. Automatic identification and locking of key targets can be achieved through intelligent sensors, monitors and biometric technologies. Using computer vision (CV) and pattern recognition technology can enable intelligence analysts to automatically extract and analyze video and voice, thereby greatly improving intelligence collection and processing capabilities.

Taking the military intelligence technical database as an example, in 2018, after conducting experimental research and demonstration, the U.S. Defense Intelligence Agency decided to develop the Machine-Assisted Analytic Rapid-Repository System (MARS) as an experimental research demonstration project. The purpose of this system is to transform the current database (US Department of Defense MIDB Comprehensive Database) that stores basic military intelligence into a multi-dimensional, flexible and rigorous data environment. MARS is ultimately designed to connect to hundreds of DoD intelligence, planning and operations-based tools and systems. The system consists of five modules, namely infrastructure, combat order, intelligence mission data, cyberspace, space, and counterspace.