Trends in AI: What’s to Come?

Legal Data and the Future of Legal Services

This post was prepared for the Law Society of Ontario Symposium on AI and its role in litigation on November 13, 2018.

The current artificial intelligence (AI) revolution is fueled by the unprecedented availability of big data. Data is the oil of the 21st century, so the saying goes. AI, algorithms and processing power are mere tools. The real value is generated from mining large data repositories. Therefore, in order to think about trends in legal AI over the next five years or so, it is crucial to start with legal data – what is it, who owns it and how does the distribution of legal data affect trends in legal AI in the near future.

The bad news for small and medium sized law firms is that other actors hold most legal data. As a result, small and medium sized firms will be consumers rather than producers of legal AI. The key challenge they face in the short to medium term will thus not be how to deploy AI, but how to develop in-house expertise in what legal technology to buy and when to outsource to legal tech intermediaries. This will coincide with a transformation of the legal eco-system as legal technology pervades and sometime replaces an ever-larger set of tasks that used to be performed by human lawyers. The future will not necessarily offer fewer employment opportunities in law, but there will be fewer lawyers working in traditional roles as legal technology becomes mainstream.

Most law firms will be consumers rather than producers of AI-based insights

AI needs data to generate insights. But where does this data come from in law? Legal data can be divided into two groups: (1) public data on the legal system and (2) private data on legal processes, files and cases.

Public legal data concerns information about primary legal sources (e.g. cases, laws, regulations), their commentary (e.g. parliamentary debates, scholarship) and actors (e.g. lawyers, judges). AI can extract value from this data by evaluating the success rate of lawyers or by predicting court outcomes, just to name a few applications. But calling this data public is somewhat of a misnomer. Whereas portals such as CanLII provide free access to court decisions to be read by humans, terms of service restrictions prevent this data to be mined by computers. Most of the “public” legal data is, in fact, in private hands – be it because the largest repositories of legal information are private companies like Westlaw or LexisNexis or because courts are unwilling or unable to share data. As a result, public legal data in large quantities is often out of reach for smaller law firms.

Private legal data is data produced in-house in the process of rendering legal services. Courts amass legal process data and law firms accumulate memos, files and filings. This data can be mined to streamline, standardize and automate legal processes. Yet, in order to meaningfully apply AI to in-house data, it has to reach a certain scale. Global law firms will be able to extract value from the data their lawyers produce in-house, but smaller law firms will not find it worthwhile to hire data scientists to apply raw AI to their smaller data sets.

In short, large quantities of public or private legal data will generally be out of reach for smaller law firms. Since AI needs data to do its magic, most law firms will thus become consumers rather than producers of insights generated by legal AI.

Navigating a growing and increasingly crowded legal technology space

As consumers of AI-based insights, most law firms will need to buy legal technology. This has three important consequences as legal technology becomes more important and replaces tasks formerly carried out by human lawyers.
First, smaller law firms will need to become technologically aware and learn to navigate an expanding and increasingly crowded legal technology space. They will need to progressively integrate new technologies into their workflows to remain competitive, provide high-quality yet cost effective services and fulfill their professional obligations. They will also face tougher procurement decisions in the process. What technology do I need to buy? How can I tell a good from a poor quality technology vendor? When do we need to buy a technology and when do we need to buy a legal technology service? To answer these questions law firms will need to increasingly invest in staff training and create new full-time positions dedicated to managing the firm’s relationship to legal technology. For smaller firms the rise of legal AI will therefore be a managerial rather than a technological challenge.

Second, a new type of law firm will emerge that will act as legal technology intermediary between legal tech and more traditional law firms. As consumers of AI, smaller law firms will not find it economical to build in-house expertise to mine legal data themselves, but they may nevertheless face a case or an opportunity that requires AI techniques. To fill a growing demand for ad hoc legal technology expertise boutique firms specialized in legal technology will emerge to support law firms on e-discovery or legal data mining matters. Often these firms will operate as local vendors and support service for larger legal technology companies. These small and often local legal technology service providers will become vital intermediaries between legal tech companies and traditional law firms.

Third, AI will spur the growth of legal technology in companies with access to large legal data. AI-powered technology will improve and proliferate to ever new legal market segments where it progressively crowds-out traditional lawyers. As the legal technology space expands and becomes more crowded, the space occupied by traditional law firms will shrink. In the aggregate, this will shift employment opportunities in the legal sector away from traditional law firms and towards legal technology intermediaries and producers of legal AI.

Not necessarily a smaller, but certainly a different profession

The uneven distribution of legal data will therefore transform the legal profession. Whereas in the longer term the disintermediation of legal services through legal technology provided directly to end-users will likely lead to a decline in legal jobs, the short to medium term will be dominated by legal technology provided to law firms resulting in a less significant effect on net employment in law. Qualifications such as legal technologists will become mainstream, firms operating as legal technology intermediaries will proliferate and incumbent and new legal technology providers will seek to mine ever-larger datasets creating new jobs for legal data scientists in the process. In short, we will have fewer lawyers working in traditional roles and more lawyers working in one way or another with legal technology.

So what should smaller law firms do to brace for this future? First, law firms need to build legal technology expertise in-house, e.g. by creating positions such as that of a legal technologist, but not overinvest in specialized skills particularly if they lack access to large legal data. Second, they will have to work more closely with legal technology intermediaries that offer specialized technology services on a needs basis. Finally, in an increasingly technology-laden world, they will have to monitor changes, threats and opportunity associated with AI’s impact on the law. Many regulatory questions from professional responsibility surrounding the use of legal technology, to the ownership and accessibility of legal data or the transparency of AI algorithms remain unresolved. These issues will be debated within the profession and in society over the next five years and will require input and active participation by lawyers.

access_time Last update September 5, 2019.

chat networking coding local-network layer menu folders diagram panel route line-chart compass search flow data-sharing search-1 message target translator candidates studying chat networking coding local-network layer menu folders diagram panel route line-chart compass search flow data-sharing search-1 message target translator candidates studying