Artificial intelligence and the law

It’s National Law Week (14-20 May) – an annual event to promote public understanding of the law and its role in society. But how is the role of the law changing? How are legal professionals adapting to new technologies such as artificial intelligence?

Dr Mark Burdon, a Senior Lecturer at The University of Queensland’s TC Beirne School of Law, sheds light on the issues that artificial intelligence raises in the legal sector and explores the new technological advances providing ‘sensor data’.

What is artificial intelligence and how does it relate to law?

When we talk about artificial intelligence, what we are really talking about is machine learning. Initially machine-learning frameworks were used to understand vast amounts of data – data that, when you put it all together, is almost unimaginable in the human context. Artificial intelligence seems to be moving towards predicting the context of the content of vast amounts of data. It is not only about understanding the types of information that can be categorised in vast amounts of data, but providing insight about how that information could be relevant to the case at hand.

In a legal context, this becomes relevant to individual cases involving millions upon millions of pages of paper which would traditionally require a large paralegal team to process the data.  What we are starting to see is the shift from human decision-making around legal consequences to machine-oriented decision-making. We then have to ask the question, ‘what’s the role of the human lawyer in the context of decisions that are going to become machine-made?’

One of your research projects at UQ is called the ‘sensor society’. Could you explain what this research is about?

The ‘sensor society’ was a project started by Mark Andrejevic and me. Mark is a professor of cultural studies in California, but was at UQ with me in 2014. We started to become interested in the development of what seemed to be quite rapid technological changes, particularly in the context of drones.

What we started to see when we looked at these different technological developments were some incredible innovations. For example, the ‘magic carpets’ developed by researchers at the University of Manchester. These carpets could identify people walking on the carpet from their unique gait patterns and analyse these patterns of data to predict when people were likely to fall. The carpets could have great social benefits in, for example, aged care facilities, to identify possible accidents before they happen.

Read more about the ‘Magic Carpet’ project here.

Behind all these developments was something quite simple, the ‘sensor’. The ‘sensor’ collects data on a 24/7 basis. It is always on and always collecting data from its environment. The amount of data that we are now generating is significantly greater than at any other time in human history. We’re not necessarily talking about the collection of individual pieces of data but about environmental collections, such as data about individuals and how individuals behave in different environments.  

What our research really focused on is that the ‘sensor’ is really just at the forefront of technological advancement. What’s happening at the back end is even more interesting.  A lot of the collected data is generated by a sensor – our mobile phones are packed with about a dozen sensors. What we are most interested in is how that data can be used to understand patterns of behaviour, patterns that we might not understand because they are so ingrained in what we do. For example, looking at a mobile phone, you can start to learn a lot about the person who uses the phone. You can now, or at least researchers are studying this, understand a person’s mood from how they use their mobile phone. If you think about the mobile phone screen, there are a number of sensors in the screen. The strength of the swipe a person uses to unlock their phone could potentially be an indicator of the mood they’re in.

This ‘sensorised data’ is becoming increasingly valuable. The real value of the sensor lies in the back end of how that data gets processed, used and stored. It’s at those points that we know far less. We don’t really understand the consequences of data analysis and data analytics that predict our behaviour. In one sense, the threat to society is about privacy. In another sense it’s something broader. It’s about the power of omnipotent organisations to understand our behaviours and how those behaviours can be targeted for certain purposes.

What are some of the new problems or issues that artificial intelligence is raising in the legal context?

Thinking about this requires us to find out in a bit more depth about what is actually happening because we really do not have a handle on many of the problems and issues. If you read some of the material that presented in the context of legal practice, there seem to be two predictions about where the legal industry is moving. At one end of the scale, it seems that the world is changing radically to the extent that law is going to be governed by machine-oriented legal practice in a very short space of time. At the other end of the scale, it seems that machine-oriented legal practice is not going to make that much difference because machines can’t replicate the human legal reasoning processes.

How will the lawyers of the future encounter artificial intelligence and in which areas of law is AI likely to be most disruptive?

To a certain extent, I’m not sure that artificial intelligence is going to change distinct areas of law. I think we need to step back and look at legal practice, legal education and the career development of young lawyers.

In the context of AI, the role of paralegals seems to be a prime focus for law firms at the moment. Anyone who’s done paralegal work will know something about the coding of documents. You have a  massive stack of documents that you have to code in a particular way so that the legal team can work with the information in the stack in a more meaningful way. Coding typically takes a team of paralegals significant time, so there’s a significant cost associated with it. One of the areas that we know machines perform better than humans is in routine tasks, so these processes are the target of increasing automation. You now have the situation where teams of dedicated paralegals (who are often law students or recent law graduates) may increasingly start to be replaced by machine-learning frameworks.

The consequences for law students and future law students are potentially profound. We have this fairly settled idea that junior lawyers do the more routine work as a sort of apprenticeship that then allows them to grow to senior lawyers. What happens when we cut that early/mid-tier out by automating this routine work? What we may be doing is removing the early- to mid-career development of future lawyers by automating the routine work they now undertake.

It’s clear that artificial intelligence is having a disruptive impact, not just on future professions but also legal learning. For us as a law school, we may have to look at changing our learning offerings in order to adapt to some of these bigger technological shifts.

Is the legal system adapted and prepared to deal with AI or are significant reforms required?

I think the answer to that is ‘we don’t know’. It’s too early to say whether significant reforms are required. What we can say is that things look like they are changing, but without a better understanding of how they are changing and what the rapid pace of change actually means, I think it’s difficult to say how the legal system should adapt. But again, as a law school that’s thinking about what our future graduates require, this is something we are looking at critically.

Are there any events or forums coming up where students can explore this topic?

We are going to be holding events related to AI throughout the year and will run a test LLB elective on AI and law over the 2018 summer semester. Significant changes are taking place. What we want to do is clarify what those changes are and, more importantly, how they are going to impact our students as future law graduates. In March this year, we organised a Q and A session about artificial intelligence and machine-learning with Professor Janet Wiles from UQ’s School of Information Technology and Electrical Engineering. On 24 May we will welcome Professor Lyria Bennett Moses from UNSW Law to discuss the technological developments in new forms of legal learning.

Find out more about the Law, Science and Technology Program at UQ.

This article was originally published on the UQ Justice and the Law Society’s blog.

Last updated:
16 May 2018