We use the following terms to describe the likelihood of a litigation outcome: “more likely than not,” “the odds are good,” or “strong case.” Rarely does a client hear “it’s a sure thing,” or “it’s a slam dunk.”

In the current information age, there are copious amounts of data available to help litigators inform their strategies and assess risk. Knowing how to use data sources and how to stay on top of emerging technology trends impacting litigation is essential for all litigators.

The Right Outcome vs. the Probable Outcome

There is a difference between the right outcome and the probable outcome, and both are important. Case law allows litigators to predict the right outcome, but predicting results is an assessment of the probable outcome. One approach assumes an ideal judge, whereas the other assumes consistent conduct. One narrows variation in facts, whereas the other assumes the relevance of broader factors. One matches to the perfect precedent, whereas the other matches to populations of data.

Probability analysis is widely used in numerous sectors, including sports, weather, medicine, and stock markets. In a business environment, for example, probability analysis can show entrepreneurs the most likely outcomes and most profitable paths.

One of the most common probability exercises that litigators use are decision trees to strategically organize and link multiple litigation issues to identify the range of potential outcomes and associated probabilities.

The Importance of Predictive Methodologies for Litigation

Litigators should become better predictors for the following reasons:

• brings numeracy to the litigation practice;
• leads to better informed settlements (influence of data on negotiations);
• reduces uncertainty and surprises for clients; and
• reduces the influence of inherent bias (confirmation bias, recency bias, sunk loss fallacy, etc.).

The characteristics of a good predictor are:

• they establish general prior probability and adjust for variables;
• they are mindful of scale;
• they revisit their predictions often; and
• they are self-critical by seeking lessons learned and avoiding easy explanations for successes and failures.

Analytics Tools for Litigation

There are several different analytics tools that litigators can use to become good predictors. These tools permit litigators to use cases to gain general insights and predictive analytics, investigate a judge’s case history, know their opponent better, and conduct more thorough research to evaluate risk and outcomes. Analytics tools that are available in North America include:

• Court Analytics (by Loom Analytics);
• BlueJ Legal;
• Context (by LexisNexis);
• DocketAlarm (by FastCase);
• Lexis+ (LexisNexis);
• Westlaw Edge (by Thomson Reuters);
• Cara (by Casetext);
• Structura (by Loom); and
• Data-Driven Decisions Program (by Lenczner Slaght), which includes the Supreme Court of Canada Leave Project, Commercial List Project, and Supreme Court of Canada Decisions Project.

Litigating in an era of “thinking” software

When transforming data into an asset, it is important to keep systemic bias in mind. Historical data used to train system can build bias into the system, which can repeat history if not used in conjunction with human discretion.

Questions that we will have to grapple with in the profession as artificial intelligence continues to gain traction are how we train articling students and lawyers to best use artificial intelligence, whether artificial intelligence can be used by the judiciary to assist with judicial decisions, whether there are any impacts on access to justice, whether there are obligations to disclose the use of machine learning (e.g., in document production), and what tasks we should be using artificial intelligence for in the place of humans.