Perceptions of Artificial Intelligence / Machine Learning in the Intelligence Community

9 Introduction Issue This study is part of a broader project to create a vision for the future of the national security intelligence mission.1 This study is a precursor to such a vision as it suggests what is known about the overall question and the current state of knowledge of problems and solutions within the Intelligence Community (IC) today. Research Question The research question is, what is the state of knowledge about problems and solutions in today’s Intelligence Community with AI/ML? Scope This study focuses on artificial intelligence and machine learning in the Intelligence Community. The terms for these technologies are often used interchangeably in practice, hence they are combined as AI/ML. Artificial intelligence can be defined as technologies that alter themselves, such as machine learning, statistical techniques using algorithms, and deep learning, a more sophisticated form of machine learning.2 Artificial intelligence can also be defined as technology that solves problems that a human typically does.3 However, there is no standard definition for artificial intelligence.4 The National Security Commission on Artificial Intelligence recently defined artificial intelligence as “ technologies that solve tasks requiring human-like perception, cognition, planning, learning, communication, or physical action; and technologies that may learn and act autonomously, whether in the form of software agents or embodied robots.”5 Machine learning benefits users based on how the technology has been trained. Supervised training means a priori data has been categorized by humans relevant to the purpose of the technology. Unsupervised training means data does not have to be categorized a priori, and deep learning uses algorithms called neural networks.6 For the most part, machine learning solves problems where a priori knowledge exists about the data complexities and how various data interact.7 Many articles use the term artificial intelligence without further specification. In contrast, other articles focus on machine learning, the most used advanced technology in the Intelligence Community. Perceptions of artificial intelligence/MACHINE LEARNING IN THE INTELLIGENCE COMMUNITY 10 Purpose The purpose for answering the research question is to ground decisionmakers in the reality of the current state of knowledge of AI/ML. This knowledge can then lead to discussions and decisions about how the IC can evolve into a more effective, relevant, and sought-after source of information so that. Relevance to the Intelligence Community This study is focused on AI/ML as it applies to the IC. The different perspectives and knowledge about this topic can provide a knowledge baseline to build ideas about its future for the IC. Research Methodology 11 Research Methodology Conceptual Framework Organizational behaviorist William Ocasio’s attention-based view of the firm theory provides the theoretical lens for this study from an individual and organizational perspective.8 Ocasio’s process-based theory is based on three premises at the individual level of analysis. First, decisionmakers make decisions based on what they attend to. Second, what decisionmakers concentrate on depends on the context of the situation. Third, the context of the situation decisionmakers find themselves in depends on the organization’s rules, resources, and social relationships. What we pay attention to influences how problems and answers are handled.

Republished by Liber Copia LLC for convenient study and reference on any device.