Data scientists from Alliance Manchester Business School are to work with a leading drug and alcohol testing laboratory on developing an automated reporting solution which incorporates advanced data-driven AI tools.
Yorkshire-based Forensic Testing Service (FTS), which tests samples for civil and criminal courts across the UK, has begun a two-year Knowledge Transfer Partnership (KTP) with the School to develop a solution that extracts and considers data from multiple sources in order to make more accurate drug and alcohol misuse recommendations for review by FTS’ team of experts.
Dr Richard Allmendinger, Senior Lecturer in Decision Sciences and lead academic on the KTP, explains the background. “This company is unique in its field in the way that it has begun to use AI in its testing processes. What is driving this is the fact that more and more pieces of evidence are being presented to courts when it comes to trying to decipher whether someone has been taking particular drugs, so the courts need to be able to distil the information quickly which AI can help do.
“We are getting to a point now where it becomes very difficult for the human overseeing these decisions to understand all the information that is being presented to them, and AI technologies can help build up a bigger picture.”
Chris Hunter, Director of Operations at FTS, said the company had ambitious growth plans which could not be achieved by simply multiplying current resources and processes.
“The preparation of technical reports incorporating evidence from multiple sources is a time-consuming and expensive process. So augmenting the tools with flexibility and constant learning will allow us to account for and report on additional types of substance misuses, such as emerging new psychoactive substances, and make informative and explainable recommendations on a case by case basis. Linking the AI tools to a wider evidence base will also give us the ability to generate evidence-based transparent interpretations so they can be understood and followed by the court.”
Hunter says the solution will provide a “step-change” in the way evidence is instructed, collected, analysed and reported.
FTS has niche expertise and experience in working with advanced analysis on a range of sample types including hair strand, blood, urine and oral fluids to assess the use of drugs and alcohol. However it lacks the expertise to develop the advanced machine learning algorithms and train them in a suitable way to produce the optimal results from the data they store and have access to, hence the link with Alliance MBS.
“We have the domain knowledge and understand the data but we need assistance to utilise this knowledge and create a sophisticated and smart algorithm which mimics the human brain neural network and deep learning capabilities to make those interpretations and decisions at a much more accurate level,” says Hunter.
Use of AI
Dr Allmendinger said the wider use of AI across the legal sector was growing. “Some law firms are now using it extensively, whereas others don’t use it at all. As part of this KTP we want to explore these issues in greater depth and will set up a user group including judges and lawyers to see how comfortable they would be in using more AI-driven data.”
However he stressed that it was crucial that AI systems were always “explainable” to the user. “The recent furore over the use of an algorithm to determine A-level results is a perfect example of the importance of explaining what a particular algorithm can and cannot do. As the use of AI grows across society this will become more and more important.”