Behavioural Analytics and UK National Security

This report presents the findings of a CETaS research project exploring the potential uses of behavioural analytics within UK national security and aims to establish an independent evidence base to inform future policy and technical innovation. The findings are based on semi-structured interviews and focus groups with participants across the national security community, other government departments and law enforcement agencies, academia and commercial organisations.

The research has found that new and emerging behavioural analytics techniques could significantly enhance investigative capability when combined with the professional judgement of human experts. However, these new use cases also present significant risks – and enhanced policy and guidance is required to ensure they can be used most effectively and responsibly within UK national security. 

Numerous areas of potential utility for behavioural analytics are identified across data discovery, inference and prediction. This includes the use of text analytics to identify author sentiment and understand behavioural patterns in online content, social network analysis for understanding roles and relationships in criminal and terrorist networks, and geospatial analysis for understanding anomalies and deviations across individuals' pattern of life which could be indicative of criminal or terrorist activity. These use cases vary considerably in terms of their levels of automation or complexity.

Beyond the technical and practical challenges in ensuring that behavioural analytics systems are fit for purpose, several ethical and legal considerations are also identified in the research. Any use of behavioural analytics that has the potential to impact on individual rights must pay due regard to the legal principles of necessity and proportionality of data collection. While these principles are already formally encoded in law, the use of behavioural analytics raises additional ethical considerations which are not explicitly accounted for within existing policy. These include the dangers of under- or over-reliance on machine outputs and the potential risks posed by the use of third-party commercial behavioural analytics products. These considerations give rise to further questions regarding transparency and public perception of national security deployments of behavioural analytics systems, which must be accounted for in future policy.

The report concludes by outlining three fictitious use cases where behavioural analytics could add significant value to existing national security processes: covering counterterrorism, serious organised crime, and insider risk detection. It is recommended that the UK government prioritises future research activity to further develop behavioural analytics capabilities across these areas with close consideration of the additional ethical, legal and policy challenges these new opportunities may pose.