Vanguard Magazine

28 JUNE/JULY 2018 www.vanguardcanada.com inteLLiGenCe AnALysis than the source having to push it to con- sumers, fundamentally changed an analyst's ability to access relevant information. In 1997, a social network service called "Six Degrees of Separation" was launched and had about 3.5 million registered users; oth- ers such as MySpace, LinkedIn, and Xing followed shortly thereafter, and by Sep- tember 2006, Facebook was available to anyone with a valid email address. Twitter, Tumblr, Pinterest, Instagram, Snapchat, Google+, and many more followed in a matter of years. Social media fed on the richness of data available on the internet and the ability to access and update that information from portable devices, such as smartphone's. Most importantly, personal data points were now also readily available to analysts needing demographic and social atmospherics information. The diagram below provides a snap- shot, as of 2007, of the Global Informa- tion Storage Capacity. Suffice it to say, the growth of data sets is at a rate that sur- passes the ability of humans to consume, interpret, and use all of the data available to us effectively. This has driven the need for advanced analytics and algorithms to manage, parse, and provide visualisation to assist with decision making. Into this mix of developing intelligence capabilities and a plethora of open source, unclassified data enters a resurgent inter- est in the development of Artificial Intel- ligence (AI), made viable by large amounts of easily accessed data and the develop- ment of neural computing. Whether ap- plying semantic reasoning or machine learning to a problem, AI provides an opportunity to significantly increase the capacity of analysts to deal with complex issues and large volumes of data. Although general AI applications are still some time from being achieved, the developments in narrow AI are only as limited as the prob- lems to be solved. In order to understand the potential of AI, I recommend viewing "Minds + Machines: A DARPA Perspec- tive on Artificial Intelligence", available on YouTube at: https://www.youtube.com/ watch?v=aWl56ohVylQ . These solutions, however, are not available "off the shelf." Each application of a given AI approach must begin with a well defined problem and strong domain knowledge. Success- ful AI start-ups consist of a strong team of entrepreneurs, domain experts, engineers, and scientists from a number of computer science and cognitive science domains. Intelligence community action Intelligence professionals now find them- selves struggling with questions of how to deal with the future security environment, a growing amount of unprocessed data that should be mined for relevant infor- mation in a timely way, and bow wave of emerging technologies that show prom- ise, but that are not easily deployed. The layered, structured data approach to sup- port the analyst has reached its limits. IPB, OPP, and traditional all source compila- tion of single source reporting is not meet- ing the mail. Capability development and training developments focus on the analyt- ical techniques and improved production processes, which is valuable, but little has been done to improve the analyst's skills in defining the problem to be solved. The ability to access data at the source by generalists has fed the desire to produce intelligence from direct source feeds, rather than from single source analysed reporting. This approach requires teams of specialists working collaboratively in a shared production space.

• Cover