Handbook home
Legal AI: Design and Development (LAWS90286)
Graduate courseworkPoints: 12.5On Campus (Parkville)
About this subject
Contact information
September
Teaching staff:
Jack Stoneman (Subject Coordinator)
For current student enquiries, contact the Law School Academic Support Office
Overview
Availability(Quotas apply) | September |
---|---|
Fees | Look up fees |
Generative AI has rapidly emerged across society, with much speculation about the opportunities and threats it poses.
As a technology that fundamentally accepts natural language instruction, has an ability to reason on that language, and provides natural language output, there are parallels with the work performed by legal practitioners— who receive client instructions, apply legal expertise with the intent to achieve some outcome, and ordinarily provide a written output in the form of a contract, an advice or a submission.
Advancements in tools that abstract away a lot of the underlying complexity mean that you no longer need to be a "software developer" to develop powerful, practical programs. Increasingly the focus is on having the domain expertise and imagination, rather than the technical skill.
To best make use of the current capabilities of this technology, and to form a view on where it might lead, there is no substitute for lifting the hood and using the core components to build practical outputs. This provides a depth of understanding that cannot be obtained by theoretical study alone.
Society will require legal practitioners who are able to understand, build, and critique the machinations of programs that use artificial intelligence (in particular, Generative AI). Accordingly, this timely and hands-on subject provides a much-needed set of tools for those engaged with law and technology: it steps through how to build Generative AI based programs for legal solutions, giving students an opportunity to put that learning into practice and become leaders in AI within their chosen career area.
Principal topics will include:
Overview of Large Language Models
o What is an LLM? How does it work?
o How can we effectively prompt engineer LLMs?
Programmatic Access to Large Language Models
o How do software programs interact with LLMs?
o How can we make LLM functionality accessible to legal users?
Information Architecture
o How do LLMs interact with data?
o What is legal data and how can it be managed?
Developing Applications
o How do users interact with LLMs – data in and out?
o How do we build applications that use LLMs?
o How can we understand and design systems that do legal work?
Measuring and Improving Performance
o What concerns should we be aware of in using LLMs?
o How can we see if an LLM is working well?
o How might LLM performance improve in the future
Intended learning outcomes
On completion of this subject, students should be able to:
- Explain the fundamental mechanisms by which a Language Model (LLM) stores information, reasons, and generates content.
- Develop a basic web-based interface that enables users to effectively interact with an LLM-based application to solve legal problems.
- Specify and implement the necessary code logic for an LLM to execute legal tasks and integrate it with a web-based user interface.
- Analyse legal documents and sources as data structures from the perspective of an LLM and demonstrate proficiency in making this data accessible for an LLM.
- Analyse the performance of an LLM-based application in solving legal problems, and enhance its performance through prompt engineering and context augmentation resulting in improved outcomes and usability.
- Hypothesise the ways in which LLM-based technology might affect the legal industry in the short-, medium- and long-term.
Generic skills
- Ability to investigate, evaluate, synthesise and apply knowledge in LLMbased AI to legal scenarios, with creativity and initiative.
- Capacity to manage competing demands on time and ability to work with a high level of autonomy and accountability.
- Capacity to value and participate in teamwork and interdisciplinary collaboration.
- Capacity to engage with issues in contemporary society.
- Advanced working skills in the design, deployment and evaluation of new technology with a focus on AI applications.
- Capacity for self-directed learning, organisation and time management.
Last updated: 8 November 2024