Solving Gun Crime With Tech
Solving Gun Crime With Tech
The way that law enforcers try to work out who committed a crime is constantly evolving to become more sophisticated. From the development of fingerprint profiling in the late 1800’s to more modern techniques such as DNA analysis and forensic testing at the scene.
But evading justice may have just become more difficult for those committing a crime using a gun with the discovery that AI can be used to determine which ammunition, and ultimately which firearm, was responsible for a particular gunshot.
Forensic scientists from King’s College London, Northumbria University, University of Lausanne and La Sapienza University of Rome have shown that this investigative work can be done using machine learning - a type of AI that can find trends in complex data. This allows experts to predict the original ‘ingredients’ of ammunition from the gunshot residue left behind on surfaces, such as spent cases, wounds, and potentially the shooter’s hands.
Previously, it would have been necessary to recreate the scenario under ‘real-life’ conditions and to carry out a test in order to make evidence ‘court worthy’. But this new method, known as quantitative profile-profile relationship (QPPR) modelling, could make the process much quicker and easier.
Dr Leon Barron from King’s explains ‘Every case is going to be different in forensic science - there are many variables to consider; different times, locations, scenarios etc. We’ve shown that despite these variables and the complexity of gunshot residue when it comes out of the end of a gun, it is possible with machine learning to drag all that information back together again to find the original ammunition used.’
In order to do this, the process also takes into account the gun that was fired, the ammunition itself and how it was dispersed, and then reads past these details.
Dr Barron says, ‘Machine learning represents one of the most promising ways to make sense of evidence more rapidly to support criminal investigations.
‘In the future, we may be able to use this technique to collect more information from the surrounding surfaces, so we can interrogate not only any ammunition used but also some individual characteristics of the person who came into contact with it. This could allow us to link the evidence together and link the evidence to the ‘who’.
‘Even now though, we use machine learning to predict an individual’s age, hair colour, eye colour, etc., just from a person’s DNA. Machine learning is being used in several exciting ways to build up a holistic picture of what has happened.’
The research team has called for the QPPR method to be applied widely in the field of forensic science and, more generally, in analytical chemistry. ‘The benefits are countless’ Dr Matteo Gallidabino from Northumbria University said. ‘They may even extend to other fields that routinely encounter changeable chemical traces, such as the analysis of improvised explosive devices, arson accelerants and environmental pollutants.’
This article has been republished from materials provided by King's College London. Note: material may have been edited for length and content. For further information, please contact the cited source.