The Law and Politics of Legal Data

• Citation: Vol. 5 No. 4
• Publication year: 2022

[Page 269]

Sarah A. Sutherland ^*

The following is an excerpt from Legal Data and Information in Practice: How Data and the Law Interact, published by Routledge.

Legal data can include many things, and there are certainly politics in organizations like law firms and professional standards for legal providers that affect how data is collected and used, but the topic of most interest from the perspective of law and politics is data associated with primary law. Whether court and legislative data is available or not and in what formats it is published is different around the world. It depends on local policy decisions, laws, and histories.

It is not possible to extricate the creation and use of data about the law from the process requirements associated with it. In early 2021, the government of Washington, D.C., was unable to pass legislation, because rules intended to allow Congress to oversee the city required the delivery of hard copies of bills before they become law. After the insurrection on January 6, fences were put around the Capital and these bills could not be delivered, so the last step to bring them into force could not be completed (Lefrak 2021).

Having legal documents available for use as data requires extensive levels of investment in publishing and infrastructure. Existing levels of investment vary greatly by jurisdiction depending on decisions governments or other organizations make, and what resources are available to deploy to ensure that legal documents are published in a way that supports various uses. In addition to these technological requirements, there are also extensive policy and legal issues that affect whether legal data is accessible and in what formats.

Beyond these requirements, law is different from many other data sources because it is constantly growing and changing. It has complex rules about whether particular rules are in effect, which are not directly dependent on time. This means that laws from any time period may need to be consulted to give a full picture of the

[Page 270]

current law, so there is more to making needed legal data available than starting a website and publishing the laws as they are passed on a go-forward basis. Instead, there is a great deal of investment involved in just obtaining the law, and even in jurisdictions that have been investing in making it available for years there are often still significant gaps that are only now being filled.

The Nature of Data-Based Applications

Data-based applications have drawn criticism for reasons such as bias against particular groups and the way they are frequently closed to scrutiny because of intellectual property interests on the part of the companies that develop them. It is important to remember that this situation is often due to political and social constraints. It is not necessarily a limitation of the technology itself, and different methodologies can support more or less transparency about how decisions are made. The issue that is almost inescapable is that machine-learning techniques in particular often use criteria for decision making that we think inappropriate or irrelevant:

The models created by machine learning are not mysterious black boxes. It is quite possible to crack them open and ask how they work. But they work so well because they are under no obligation to condense the world into a crisp definition. A model of a complex, over determined concept is likely to be just as over determined itself. (Underwood 2019, 49)

The adoption of tools that cannot be adequately examined for high stakes applications, especially those used by states, will need to be thoroughly examined for fairness if they are to be sustainable solutions. A possible solution is to only use open source software and solutions for these applications to allow for scrutiny of the criteria used and outcomes.

It will be interesting to see how this will move forward. It will be important not only for political and social concerns, but also for the validity of findings: "Because the relationships relied upon in Big Data are entirely empirical and both concurrent and predictive validity are time dependent . . . there is no reason the correlations that underlie Big Data solutions should persist beyond the sample period" (Crews 2018, 94).

[Page 271]

Access to Data

Even where the law is available in electronic formats, it is frequently not accessible for use in application development and research to the degree that those looking to use it may wish it to be. Many countries' legal publishing traditions started from a background of private publishers making volumes available for sale when governments did not. In common law countries, the most frequent model for the start of publishing case law was for private interests to start publishing the law because governments had no interest in doing it.

Issues like this have continued to the present day in many parts of the world, with concerns like disagreements about the intellectual property status of primary law and limits on access to documents being ongoing in many jurisdictions. This leads to situations like the litigation in Georgia v. Public.Resource.Org., Inc. where Carl Malamud was accused of terrorism for publishing the Code of the American state of Georgia in 2019 after the state had entered into an agreement with a private publisher to publish the state Code and provide annotations. The Code itself was not protected by copyright, but the annotations were (Liptak 2019). The United States Supreme Court found in Malamud's favor.

This complexity leads to private and corporate interests in public data that will be difficult to extricate, illustrating the tension between corporate investment and profits and public spending and interests that have been ongoing in many parts of the world. There does not seem to be a great push among governments and courts outside a few jurisdictions, such as the United States, the European Union, Singapore, and recently the United Kingdom, to make law widely available for private use as data sources, which would allow easier access for exploration of new uses for the data.

In many countries the main publishers of primary law continue to be outside government, with non-profit organizations such as CanLII in Canada, AustLII in Australia, and BAILII in the United Kingdom, and a combination of private publishers and non-profit organizations filling that role in the United States. The fact that in many cases the issuing governments have not made the law available for wider development, and, in fact, may not have preserved copies of the documents in a way that would allow for further distribution means that these organizations have faced pressures

[Page 272]

...