In my previous post, I discussed lawyers, and litigators in particular, as analysing complex data sets to provide clients with reliable predictions and risk-balanced strategies. As such, their processes are fertile ground for predictive analytics and machine learning. Here I’ll discuss some categories of use cases, with powerful social benefits.
- Disputes outside the formal litigation process
The first category of use case will address disputes outside of the formal litigation process. An enterprise version focuses on pre-litigation claim management. These could tackle single-issue questions on a non-precedential basis. Early examples will help enable legal questions to be more efficiently resolved by applying a set of rules to be followed in problem-solving operations, that is, an algorithm. Early examples could include an appeal system operated by a local planning or transport authority; a traffic-fine appeal system could even arguably support itself, charging a small fee to apply for the fast track process which could be returned if the appeal is successful. Reducing their decision systems to an algorithm would ensure authorities are transparent and rigorous. With more disputes are resolved without needing to reach litigation, let alone the courtroom, legal inclusion will rise.
View our infographic on predictive analytics
- Preparing for litigation
The second category of use case will assist trial lawyers in preparing for the litigation process. Predictive analytics tools will help trial lawyers to validate litigants’ claims, and by detecting judicial patterns, they will ensure more consistency in legal assessment, providing more certainty. Firms are already partnering to develop great tools to help expedite legal research. These will develop into powerful predictive analytics which will help lawyers ensure sub-optimum cases are settled earlier – or avoided altogether, considering factors such as the likelihood of ever collecting any award based on the credit profile of the opposite party. The next wave of apps will help uncover insights to help shape legal strategies: for example, does this judge tend to rely on certain precedents, or express particular criticisms of certain decisions?
- Manage the litigation process
A third category of use case will help judges to manage the litigation process. Fortunately, judges are selected from a pool of the most successful trial lawyers; they are used to distinguishing between causing harm and being financially and legally liable for it. However, the physical court process and manual analysis do require time. Some may hesitate, perhaps preferring unfettered human judgment (although inevitably inefficient and inherently flawed), but is that in our best interests? Judges currently rely on clerks to conduct preliminary online research and prepare drafts, so their frame of reference is already bound by such reliance. Superior software means judges have the same powerful tools as the lawyers appearing before them, and can search more for themselves.
- Enforcement
A fourth category of use case will assist the enforcement of the results of the litigation process. Whilst a court may awards damages, it may be practically difficult to collect these from the debtor if they refuse, or claim they cannot afford to pay. In some cases, an order garnishing wages is appropriate; however, it may only be possible to collect a meaningful proportion over a long period of time. A system of automatically updating multiple company, property or credit databases with adverse judgements could assist collection by flagging when income or assets become available or accessible in the future.
- Compliance requirements
The final category of use case will assist with compliance requirements in regulated markets. Financial data scientists are increasingly turning to financial regulation to assist regulators with risk monitoring, surveillance and supervisory capabilities and regulatees with the sheer volume of compliance obligations. In the financial markets, complex trading analysis and massive execution activity is conducted in milliseconds by robo-advisors using low latency connections. When highly complex financial analysis applied at scale is left to machines, only similar machines may be capable of providing regulatory oversight and helping to guard against systemic risks.
The key insight that drives all of these types of predictive analytics is that when you turn words into numbers you can create scalable relationships, add those up, and get the data to reveal useful things. By surfacing insights from legal decisions it could help prevent errors and even miscarriages of justice. Unusual trends can be flagged and examined. They could be trained to ignore rhetoric and weight certain public policy objectives more accurately and consistently. It may also help inform the drafting of regulations, as well the design of compliance training. If algorithms are already helping the best doctors detect and treat cancers, using them to assist judges in administering justice must be welcomed if it means we end up with, on average, better decisions.
