Tuesday, April 26, 2011

Legal Technologies that Scream for Improvement

When anyone with a technical background-- in my case computational neuroscience --enters the legal profession, it's like Dorothy going back to Kansas.  There is no more Technicolor: everything becomes gray and somewhat dated.  It's like a time-warp.

Now, some of this is natural and a result of the position that law has in society, as a gatekeeper and protector of existing traditions.  But much of it is simply frustrating and, I think, one of the reasons that so many lawyers are unhappy in the profession.  What makes it more frustrating is the potential that law provides for creating structures that "play well" with computers: like computer programming, it is an all-text medium, where units are (or should be) logically related and interlinked.  Just as a programmer can write unit tests for computer programs, lawyers should be able to use technology to provide more rigorous legal analysis and to make legal processes more efficient.

As a starting point to tackle what can be done about the gap between potential and actual legal technology, I'll list some of the major areas that cry out for improvement:

1. Access to Basic Legal Information

Legal information is fragmented and generally not well organized in electronic form.  All primary legal information is legally in the public domain, but in practice may be hard to find or in a form that makes it impossible to search electronically.  While lawyers at large law firms have an advantage, with access to expensive databases of legal information curated by thousands of staff attorneys at the legal publishers (Westlaw, Lexis, etc.), the fragmentation of data always leaves a fear that the information is incomplete. And even as primary legal materials become available, Professor Richard Leiter and others have pointed out that lawyers will continue to rely on the commentary of their peers and other secondary sources in order to make sense of the law. A "Free Access to Law" movement has gained momentum, especially in the UK, as the tools to "crowdsource" or automate collection of primary legal information have matured. But we're still far from the goal of open and accessible legal information.

2. Research, Writing and Collaboration Tools

Lawyers primarily rely on Microsoft Word to write their documents, and often transmit them by Adobe pdf to the court and other parties.  These proprietary formats are used largely for layout and, in the case of pdf, security.  But they also make it far more difficult to apply standard computer science tools to create and analyze documents, which means that lawyers end up doing a lot of things by hand that programmers let their editing environments do for them automatically.  Programmers take these tools for granted and don't understand what's the big deal about building such tools for lawyers; when I ask them imagine writing their code in MS Word, they get it.

3. Legal Analysis Tools

In principle, law is well structured to apply semantic analysis techniques, to determine important legal precedents, and even to automate tasks like determining whether a given statute or regulation is current. This area again, is one where the computational tools were developed long ago, but where legal information is not produced in a clean or consistent format, making such analysis many times more complicated.

4. Mentorship in the Legal Profession

It may seem strange to list mentorship among the areas that need technical improvement.  However, major job losses in the major firms and changes in their hiring patterns mean that more lawyers are starting out on their own.  While many will leave the legal profession entirely, this also creates the potential for more "virtual" law firms.  Technological platforms for collaboration and mentorship will need to develop in order to support this change.

5. Consistency in Court Procedure and Data Formats

The Federal courts are all required to publish their legal opinions online.  But a brief survey of their websites shows that each (while looking like a '90s throwback) organizes the information of the court differently, offers different formats and, in general, needs to be dealt with individually.  This is even more true at the state level, creating a fragmented system of information that can, currently, only be approached with the large budgets of the major publishing companies.  The potential for automation is obvious, and there are open source projects (e.g. here) to collect and publish Federal court opinions, but the lack of format consistency is a major barrier.

6. Consistency in Legislation

A similar inconsistency is found in the publication of legislative information.  Distributed, open source efforts, such as the Sunlight-Foundation-sponsored Fifty State Project are trying to tackle this at the state level, and others, such as Cornell's Legal Information Institute, are working with Congressional offices to help structure and hopefully standardize Federal legislative information.

These are just a taste of the fundamental challenges to bringing order to legal information and legal practice.  Other areas, such as writing and analyzing contracts, could also use major overhaul. One common thread is the poor quality of much of the legal data that is being produced.  The best place to create structure in legal data is when it is first produced: when a law, regulation, memo, brief or opinion is drafted.  The best technological tools can only do so much with data that is a mess: GIGO. More work, clearly, needs to go into creating cleaner, more structured legal information.  Two complementary pieces are needed to make this happen: use of plain language in drafting, and the use of drafting tools (e.g. this) that impose some structure on legal information from the outset.