AI as the Sixth Intelligence Discipline: A Tradecraft Guide
I wrote this guide to prepare intelligence analysts for the future of intel analysis. It introduces analysts to foundational concepts to understand AI as a new source of intelligence. From that they will be able to delve into the analytical tradecraft, from the technical knowledge needed to evaluate models to how to fuse their outputs with other traditional sources of intelligence, and much more!
I
Why this Guide?
Like Simon Sinek advises always start with "why." In this chapter, I detail the reasons I believe this guide is necessary and significant. First, it introduces and establishes the groundwork for a new intelligence tradecraft. As with anything novel, my hope is that this will propel others to contribute to and expand upon it. Second, it aims to prepare current and future intelligence analysts for what I perceive to be the future of intelligence analysis. AI is pervasive, and developing a specialized understanding of its application to intelligence analysis is crucial. Ultimately, these two reasons underpin the third and most vital one: aiding the US and its allies in maintaining and enhancing their decision-making advantage and information superiority.
II
AI as a Peer Collection Discipline
This chapter treats AI as a counterpart to other collection disciplines such as HUMINT, SIGINT, GEOINT, and OSINT. It delves deeply into essential AI concepts and themes that underscore this peer-to-peer relationship. The chapter is pivotal in positioning AI as information rather than just a tool, which is a central theme of the book. This entails that AI, like other collection disciplines, is subject to the collection and analysis processes.
III
AI as the Sixth Intelligence Discipline (AI-INT)
This is the most critical and significant chapter of the book. Here, I explain why and how AI should be regarded and conceptualized as the sixth intelligence discipline, which I term AI-INT. I describe how AI-INT fits within the intelligence cycle, the process of creating AI-INT models, and answer key questions concerning AI-INT models. It includes a breakdown and example of AI-INT models on Python to solidify the key concepts.
IV
AI-INT Analytical Tradecraft
In this chapter, I dissect how intelligence analysts can integrate AI-INT with other intelligence sources and establish the foundations of the tradecraft. I begin by discussing how analysts can assess AI-INT through an access, credibility, and reliability framework; how they can incorporate both raw and disseminated AI-INT into analytical processes and methodologies; and how they can write it in finished intelligence products.
V
Future of AI-INT Tradecraft
In this chapter, I explore the current deficiencies of the tradecraft, how we might start addressing these gaps, and its potential trajectory over the next five years.
VI
The Ten Principles of AI-INT
When I began in government, a senior analyst emailed me the "golden rules" for analysts—ten principles that every analyst informally aspires to uphold. I do the same here with AI, AI-INT, and its role in intelligence analysis.
VII
The First Lego
The title of this chapter reflects my view that this book is the initial building block—like the first Lego in a set. In it I encourage others to add more pieces to eventually construct something significant. This chapter is an invitation and rationale for others to build upon, enhance, and develop new concepts to further advance this tradecraft.
