Sunday 30 October 2016

Using Artificial Intelligence for proactive defense

We need AI to understand code and applications. Our code complexity is getting to a level that we need to start to use artificial intelligence capabilities to understand it, and to get a grasp of what is going on, so we can create secure applications that have no unintended side effects.

As AI becomes much more commonplace, we should start to use it more to source code analysis and application analysis. Kevin Kelly has some very interesting analysis on the use of AI, where he discusses the idea that one of the next major revolutions will be where we start adding AI to everything, because the cost of AI will become so low that we will be able to add AI to many devices.

When you analyse an app, you should use everything you have. You should use static, dynamic, interactive, human, and increasingly you should use artificial intelligence to optimise your analysis.

When you are doing security analysis, you are dealing with a vast amount of data, displayed on a multi-dimensional graph. What you have is a graph of the relationships, of what is happening. You are looking for the connections, for the paths within the graph, that are made of what is really going on and what is possible.

Artificial intelligence technology can assist the human who will put context on those connections. I think we are a long way from being able to do this kind of analysis automatically, but if we can make the human's job of reviewing the results easier, or even possible, that is a major step forward.


(from SecDevOps Risk Workflow book, please provide feedback as an GitHub issue)