Actionable open-source intelligence for ‘cold cases’
The number of ‘cold cases’ in the Netherlands is rising every year, with more than 1700 in total right now. To gather new evidence Swikar Bhandari is researching the capabilities of open-source intelligence. ‘The proposed system will include the capabilities of both artificial intelligence (AI) and citizen science-based approaches,’ he says. ‘In order to derive actionable intelligence.’
Use of open-source intelligence (Osint) in high-risk domains is growing to be more common. This includes sectors of law enforcement and military, for applications such as identifying terrorist threats and fraud detection. ‘The full potential of Osint for homicide investigations, in particular cold cases, is still an unexplored territory,’ says Swikar Bhandari. ‘But a few successful examples in applications do exist.’
A good example is forensic genetic genealogy. DNA databases developed, using genetic data of citizen volunteers, have helped to identify missing and unidentified people as well as perpetrators who previously managed to get away with their crimes.
Combining diverse sources
PhD-candidate Swikar Bhandari, member of the University of Twente (UT) group Data Management and Biometrics (DMB), stresses that intelligence derived from open-sources such as the internet cannot be trusted at face-value. Serious problems arise due to the reliability, inconsistency and ambiguity of the various data.
In his quest to construct a reliable Osint-based system, Swikar implements the concept of probabilistic data integration. This helps to manage the uncertainty of the thus derived information. He additionally explores the concept of crowdsourcing or citizen science. Swikar: ‘This is a new socio-technical approach. The goal is to design a system that facilitates citizens' cooperation in investigations and guides them in collecting and evaluating information to obtain useful (actionable) information.’
Methodological ways
Swikar and his research fellows hope to design a semi-automated system that can potentially re-open cold cases and may help investigators to identify the perpetrators of these heinous crimes. The system will aid decision-makers in their search and reasoning activities to gather new evidence.
In order to ensure safety, security and responsible use of decision-support systems, Swikar is exploring new methodological ways to evaluate the reliability of open-source data. ‘Ethics comes into play here,’ he says. ‘Ethical and legal implications of collecting open-source data in the derivation of actionable intelligence remain fundamental issues.’
Multidisciplinary approach
Because of the multidisciplinary nature of this project, Swikar works together with the Philosophy section of the Technology, Policy and Science (TPS) department at the UT-faculty of Behavioural Management and Social Sciences. Also the Technologies for Criminal Investigations research group of Saxion Hogeschool and the Police Academy of the Netherlands in Apeldoorn are affiliated.
‘Technological and philosophical perspectives are both part of my scientific output,’ Swikar says. He presented his first research output at the Homicide Research Working group 2024 annual conference in Florida in June, based on the theme: the use of technology and other innovative methods. His paper explained how the accommodation of solvability factors can enhance the data quality of homicide databases.
Motivation
‘Now I work towards validating existing data collection methods through a comparative analysis of narrative approaches for homicide investigations,’ Swikar says. Since a young age he was fascinated with the idea of using technologies for criminal investigations. After obtaining a master’s degree, he explored PhD projects that fulfilled this interest. ‘The project at the UT suited my research interest unbelievably well.’
