A new report published this week by The Alan Turing Institute's Centre for Emerging Technology and Security (CETaS) explores the use of machine learning for intelligence analysis, based on in-depth consultation with stakeholders from across the national security community.
The research found that machine learning could help analysts to sift through large amounts of data, helping them to prioritise the information that matters most to help keep people safe. The ongoing global expansion of data presents both risks, and also opportunities. The use of ML offers real potential to simultaneously reduce such risks and to pursue such opportunities.
Machine learning works by identifying connections and patterns across large volumes of data that may otherwise go unnoticed by human operators.
The report offers fresh insight into how intelligence analysts make decisions, and it offers a series of recommendations for the design and deployment of machine learning models to ensure they could be effectively and responsibly adopted within national security.
The CETaS researchers found that in order to use ML effectively, those implementing the technology need to have a good understanding of the analyst’s real-world work environment. And while improving the analyst’s understanding of how ML works is important, the researchers report that it is not enough for analysts to trust the technology.
The researchers recommend that explanations on the use of ML should be tailored depending on the person’s expertise. For example, data scientists or those developing policy or approving the deployment of a ML system require different explanations to those required by analysts and oversight bodies. This is important because current efforts to explain how ML works focus on mathematical explanations that day-to-day users may not understand and do not help reinforce trust in the model.
This report is one of the first public studies investigating human-machine teaming and the use of machine learning for intelligence analysis within UK national security.
Anna Knack, lead author and Senior Research Associate at The Alan Turing Institute, said: “It’s time consuming for intelligence analysts to work through the vast quantities of data that come across their desks every day. Machine learning can speed up this process, filtering out the irrelevant information helping analysts to act on information much more quickly.”
The use of machine learning to support intelligence analysis presents new challenges such as how to present the right amount of technical information regarding the model's performance to the user, to ensure they maintain the appropriate level of trust in the system. The recommendations in this report are aimed at ensuring the responsible, effective and proportionate use of machine learning for intelligence analysis so national security agencies can effectively deploy this technology to help keep people safe.
