A Vision of a Toolbox for Intelligence Production Joel Brynielsson, Andreas Horndahl, Lisa Kaati, Christian M˚ artenson, Pontus Svenson Swedish Defence Research Agency SE-164 90 Stockholm, Sweden Email: firstname.lastname@foi.se Abstract—In this paper, we describe preliminary work on a toolbox aiming to help analysts involved in the intelligence production process. Intelligence analysts are overwhelmed by information, both in the form of sensory data, text stemming from human observations and other sources. In order to make sense of this information and to produce the intelligence reports needed by decision-makers, assisting computer tools are needed. We briefly describe parts of the intelligence process and touch upon the subject of what parts can and cannot be automated. A tool for tagging information semantically that we are currently working on is described, and ideas for two other tools are briefly outlined. I. I NTRODUCTION Intelligence analysts of today are overwhelmed by infor- mation that they must take into account when producing their analyses and assessments. They must also produce far more content today than previously while at the same time taking into account material from a wide variety of sources, ranging from IMINT (image intelligence) to text articles in newspapers. Both the pace of the information push to the analysts and the information pull by the commanders and decision-makers who need output from the analysts has also increased: the time-scales involved at operative and tactic levels are shorter compared to the days of the cold war. The types of conflicts that we are involved in have also changed: in peace-keeping and peace-enforcing missions, we are faced with a multitude of actors, who are of different types and it is not always clear whether they should be regarded as friends, foes or neutrals in a given situation. When analyzing, for instance, a report on a confrontation between two clan leaders, it is important not only to know the interests, capabilities and motivations of these two, but also those of the other actors in the area who are connected to the involved parties. The goals of network-based defense and network-based intelligence will further exacerbate these problems: more sensors mean more data for the analysts to consider, and also more decisions to be made regarding where to put the sensors. The need for continuous action can make the traditional intelligence cycle obsolete. Instead, it is necessary to make all stakeholders, including customers, involved in the intelligence process, en- abling a creative collaboration process where several analysts may contribute in parallel to a continuously refined shared picture of the target, see, e.g., [1] who proposes a “target- centric” intelligence cycle. In order to meet these requirements, intelligence analysts need better information handling tools and concepts. Fusion tools can be an important help, enabling automatic clustering of similar reports and using aggregation methods to produce meaningful labels on the information displayed to them. Techniques from natural language processing and text mining can be used to fuse information in different languages and to produce summaries of vast amounts of textual data. Complete automation of the intelligence production process, however, is neither possible nor desirable. Hence, the task for fusion researchers is both to build automatic tools that process information and to build tools that help humans to do further processing: ultimately, fusion is a process involving humans. Situation awareness is created in human minds, not in machines. In this paper, we present some ideas on how computer tools for intelligence analysis can be created, i.e., tools that can be used at the discretion of a human analyst to produce intelligence products in alternative ways. What is needed, rather than constructing one single tool, is a “toolbox” containing several different tools, each helping the analyst with one specific task. The work described in this paper emphasizes intelligence analysis, i.e., the work of intelligence analysts which can best be described as the art of creating useful intelligence products. At the same time, however, the work described herein is strongly related to work within other areas currently performed at the Swedish Defence Research Agency (FOI) in various ways. All the ideas described in this paper will eventually be included in the Impactorium toolset which is described further in [2]. A future version of the semantic tagging tool described in Section III will be used to help users input information into the Reportorium tool of the Impactorium suite as well as to input information into the Semantic MilWiki [3]. The threat model construction tool described in Section IV-A will be used to construct the threat models used in Impactorium, and one of the uses of the need-based situation picture tool described in Section IV-B is to determine which results from the Impactorium and what information from the Semantic MilWiki should be displayed to what users at what time. This paper describes ongoing research and does not present finished research or even finished thoughts about further research. Section II contains a description of different parts of intelligence analysis work where it could be possible to use computer-aided tools. Section III contains some details about a tool for semantically tagging information that we are currently working on, while Section IV gives a brief overview of envisioned tools for constructing threat models and need- based situation pictures. Finally, Section V concludes and discusses future work.