National insurance law firm BLM has teamed up with three professors from the London School of Economics (LSE) in a two-year research project to create models that predict the cost, length and outcome of litigation.
Andrew Dunkley, head of analytics at BLM, said artificial intelligence (AI) would be an “important part” of the partnership, but the crucial thing was the way it blended technology with actuarial knowledge and ‘decision science’.
“What decision science looks at is taking decisions in an uncertain environment, like litigation, where you don’t have complete information,” Mr Dunkley said. “You may not have enough information to come to a confident prediction.
“At the start of the dispute, you have a bit of information about the case, but there is lots of information you don’t yet know. You still have to make a decision about whether or not to invest in the litigation.”
On the LSE side, the partnership will be led by Professor Henry Wynn, head of the decision support and risk group. He will be joined by Professor Pauline Barrieu, head of LSE’s statistics department and principal examiner for the Institute of Actuaries, and Professor Milan Vojnovic, chair in data science and expert on machine learning.
They will be supported by a dedicated post-doctorate fellow who will work with BLM’s analytics team. The project will be co-funded by BLM and LSE’s Knowledge Exchange and Impact Fund.
BLM set up its specialist data analytics team last July with the aim of helping its lawyers make better predictions about their cases.
Mr Dunkley said the focus would be on both high-volume claims and low-volume, high-value claims.
He said BLM had a “huge amount of data” on high volume claims, where AI could be used to help triage and value them. With higher value claims, ways could also be found to help lawyers make better predictions and clients manage their risk.
“It can help in deciding whether to engage in litigation, whether to settle or not and how to resource a claim successfully. Accurate resourcing is a really important issue for our clients. Every pound held in reserve is money that cannot be invested in the market.”
Mr Dunkley said improvements in forecasting could encourage earlier settlements, and as things improved further, costs may come down.
He said that in those parts of the legal services market dealing with high volumes, whether of documents or cases, AI was “moving from ideas to things that will change the way the industry works”.
He predicted that, in 10 to 15 years from now, this part of the industry would “not look the same as it does today and it will be driven by AI.”
Mr Dunkley added that in “any other industry” outside professional services, research and development partnerships were perfectly normal.
“We want to be driving the conversation, rather than at the mercy of it.”
Professor Wynn said the LSE statistics department and its decision support and risk group had a “substantial interest” in quantifying risk and uncertainty.
“Building up a joint centre of research excellence in the valuation and management of litigation risk with BLM is an exciting development in this area, with a clear commitment on our part in terms of the investment and funding of this work.
“We look forward to working together to develop BLM’s ability to forecast the outcome, cost and length of litigation, and to understand what drives these factors.”
There is a small but growing number of such knowledge transfer partnerships. Last November, fellow defendant firm Kennedys announced a two-year collaboration with AI experts at Manchester University to develop a new product to combat insurance fraud.
Southport-based Fletchers Solicitors, which handles claimant clinical negligence and serious injury work, has tied up with Liverpool University to develop AI technology to support decision-making within the firm.
Pioneering alternative business structure Riverview Law went down this road three years ago when it also teamed up with Liverpool University to leverage the university’s leading AI expertise.
In 2014, London firm Hodge Jones & Allen teamed up with a leading academic to pioneer a predictive model of case outcomes to enable the firm to better assess the viability of its personal injury caseload.