PI Brief Update

How AI is Changing the Legal Landscape - Dr Mark Burgin

24/05/23. Dr Mark Burgin explains the different AI technologies available and how they will and are being used to disrupt legal practice.

When a legal practice considers whether AI will assist their business it is easy to become overwhelmed. AI is cheaper and has had recent a step change in effectiveness with AI like ChatGPT so it cannot be ignored any more. There are so many different types of AI technologies available, each with their own promise. In this article I describe the main types and what they are good at.

Which AI system is most useful depends upon the number of hours the practice spends on each task. Practices can choose to improve the AI further by using their own dataset to train an AI system. Typically AI can increase an individual lawyer’s output by 4 times but some lawyers will find the technology easier to use than others. The cost of training the AI is a fixed cost so the more hours spent on a task the quicker the money could be recouped.

It is important to say that AI reliability is improving but it is better at generating options for a human to select than finding a definitive answer. Some lawyers will prefer to iterate organically working to a better option. Other lawyers will get the best results by asking AI to give reasoning for each of its choices. The aim should be for man and machine to work together rather than have AI make the decisions.

Options for AI systems.

• Machine learning: Machine learning AI systems can be used to review large volumes of documents, such as contracts and legal briefs, to identify relevant information and potential risks. For example, a machine learning AI system could be trained on a dataset of contracts to learn how to identify clauses that are likely to be disputed. Once the system is trained, it can be used to review new contracts to identify potential risks.

• Rule-based AI: Rule-based AI systems, on the other hand, are programmed with a set of rules that they use to make decisions. For example, a rule-based AI system could be programmed with the rules of evidence to for instance assess the compliance of expert reports with the procedure rules.

• Neural networks: Neural networks are a type of machine learning AI system that is inspired by the human brain. Neural networks are made up of many interconnected nodes, each of which represents a different feature of the data that the system is trained on. For example, a neural network could be trained on a dataset of legal cases to screen new cases and predict their outcomes.

• Reinforcement learning: Reinforcement learning is a type of machine learning that allows AI agents to learn how to behave in an environment by trial and error. The agent is given a goal, and it is allowed to take actions in the environment. The agent is then rewarded or penalized for its actions, and it uses this feedback to learn how to achieve its goal. For example, a reinforcement learning AI agent could be used to learn how to negotiate a settlement in a legal case or perform legal research.

Negotiation: AI can be used to help lawyers negotiate more effectively. For example, AI can be used to identify the strengths and weaknesses of each party's position, and to suggest possible compromises. For example, an AI system could be used to help lawyers negotiate a settlement in a legal case.

Legal research: AI can be used to research legal precedent and case law to find relevant cases and rulings that have a good fit for the current case. For example, an AI system could be used to search for cases that are similar to the current case in terms of the facts, the legal issues, and the outcome.

• Natural language processing (NLP): NLP is a field of computer science that uses transformers which work on the boundary between computers and human (natural) languages. Transformers are made up of self-attention layers, which allow the network to learn long-range dependencies in the input data.  NLP is used to understand and generate human language, and they are used in a variety of applications, such as machine translation, speech recognition, and text summarization. For example, an NLP system could be used to summarize a legal opinion or to translate a legal document into another language or for Dispute resolution.

Dispute resolution: AI can be used to resolve disputes more quickly and efficiently. For example, AI can be used to mediate disputes between parties by generating replies, or to create binding arbitration awards.

• Generative adversarial networks (GANs): GANs are a type of deep learning model that can be used to generate realistic images, text, and other data. GANs are made up of two neural networks: a generator and a discriminator. The generator is responsible for creating new data, and the discriminator is responsible for distinguishing between real data and generated data. The two networks compete with each other, and this competition helps the generator to learn how to create more realistic data. For example, a GAN could be used to generate real world solutions to problems that do not have a simple answer.

• Deep reinforcement learning: Deep reinforcement learning is a combination of reinforcement learning and deep learning. Deep reinforcement learning systems are able to learn complex policies by interacting with an environment. They are used in a variety of applications, such as game playing and robotics. For example, a deep reinforcement learning AI system learned how to play chess so could navigate a complex legal environment. Artificial intelligence (AI) is rapidly changing the legal landscape.

Conclusions

Artificial intelligence (AI) is rapidly changing the legal landscape. AI is being used to automate tasks that were once done by lawyers, such as legal research and document review. AI is also being used to develop new legal tools and services, such as predictive analytics and legal chatbots.

When considering which type of AI to buy and train it is important to consider where time savings can be made. Tasks that are time-consuming and repetitive, such as legal research and document review are better than more complex tasks, such as strategic planning and client counselling.

AI is becoming more affordable and accessible so that even small and mid-size law firms will benefit. A law firm could rapidly expand using AI to reduce costs but there are also risks. AI can be expensive to implement and maintain. Second, AI can be complex and difficult to use. Third, AI can be biased, and it can make mistakes.

Choosing the right AI for the right task can transform productivity but the wrong AI will drain the business of time and money. Equally failing to adopt AI will leave a legal practice facing increasing workloads for less and less money. Many lawyers are already feeling the effects of AI on their area of the industry such as PI.

The first step is to have an informal conversation with an amateur systems analyst who has worked in the legal sphere for a decade or two. Communicating with an AI expert is expensive and difficult. The practice needs to understand their options and whether there are cheap over the counter solutions already available. If it necessary to create a bespoke system, then the practice will already be up to speed on what they want to achieve and the likely costs.

This article was written with assistance from BARD.

Doctor Mark Burgin, BM BCh (oxon) MRCGP is on the General Practitioner Specialist Register.

Dr. Burgin can be contacted on This email address is being protected from spambots. You need JavaScript enabled to view it. and 0845 331 3304 website drmarkburgin.co.uk

Image ©iStockphoto.com/Just_Super

This is part of a series of articles by Dr. Mark Burgin. The opinions expressed in this article are the author's own, not those of Law Brief Publishing Ltd, and are not necessarily commensurate with general legal or medico-legal expert consensus of opinion and/or literature. Any medical content is not exhaustive but at a level for the non-medical reader to understand.