ChatGPT - Shifting Ground for Junior Lawyers

Artificial intelligence (AI) is rapidly transforming many industries, including the legal profession. As AI technologies continue to advance, they are likely to play an increasingly important role in the work of lawyers. AI systems can assist lawyers with a wide range of tasks, from conducting legal research and analysing large amounts of data to drafting documents and identifying patterns. By automating routine and time-consuming tasks, AI can help lawyers become more efficient and effective in their roles. However, the increasing use of AI in the legal profession also raises concerns, including the potential for job displacement and bias in AI systems. As these technologies continue to evolve, it will be important for lawyers to understand and adapt to the shift.

1. Introduction

The paragraph above was crafted by ChatGPT using the prompt “Write an introduction to the role of artificial intelligence and the effect it may have on junior lawyers going forward.” By utilising advanced machine learning techniques, ChatGPT is able to generate text that closely resembles human writing, providing summaries, explanations, and even original content such as poems and music. It may yet revolutionise how we access and understand complex ideas drawn from a wide range of sources.

With the ability to automate a wide range of tasks and provide instant access to vast amounts of information, AI has the potential to materially change the way legal work is performed. That raises obvious questions about job prospects and career development for junior lawyers, as well as the possibility that AI will fundamentally alter the texture of legal work itself. Tools like ChatGPT may lead to a shift in how junior lawyers are used, with greater emphasis placed on higher-level tasks, judgment, and decision-making.

2. What Is ChatGPT?

ChatGPT is a large language model built using a type of artificial intelligence called a neural network. Neural networks are loosely inspired by the way the human brain processes information. In that structure, individual units, or “neurons”, are connected to one another in layers, with each neuron receiving input from multiple other neurons and passing its own output onward.

The way a neural network “learns” is through training. This involves feeding the network large amounts of data and adjusting the strength of the connections between neurons based on how well the network performs a given task. In the case of ChatGPT, the network is trained on large amounts of text, and those connections are adjusted based on how well the model can generate coherent and natural-sounding language.

Through that training process, the model learns to recognise patterns in data and generate text from those patterns. That allows it to perform a wide range of language-related tasks, such as translation, summarisation, and conversational response generation. The underlying applications are already familiar: chatbots, drafting assistants, machine translation, and content generation systems.

The history of OpenAI goes back to 2015, when it was founded to advance artificial intelligence and encourage its responsible development. Since then, it has become one of the most visible organisations in the field. The GPT line of models is designed to generate human-like text by predicting likely continuations in a sequence of words, and GPT-3 became the best-known version because it demonstrated just how fluent that approach could feel in ordinary use.

3. A Practical Reality

Law students and junior practitioners have long been exposed to buzzwords and concepts related to disruption, particularly in legal technology and the changing nature of the profession. For many, those ideas remained largely theoretical, with little practical relevance to everyday study or practice.

Technologies like ChatGPT change that dynamic. Law graduates are now likely to encounter AI tools in their near-future careers and will need to be prepared to adapt. In that sense, ChatGPT represents the practical application of disruption concepts that law students have spent years debating. What once felt abstract is now operational.

One of the clearest applications is document drafting. By providing the chatbot with legal context and terminology, it can generate drafts of contracts, legal briefs, and pleadings. It can also be used to provide information and guidance on legal questions, such as answering queries about laws or regulations.

As AI tools become more advanced and able to handle a broader range of tasks, there is clear potential for some work traditionally assigned to junior lawyers to be automated. As routine or repetitive tasks are taken over by software, junior lawyers may be pushed toward work that requires higher-level thinking and problem-solving. They may also find themselves working alongside these technologies every day. That could reshape the way legal work is organised, as well as the way junior lawyers are trained and supported.

The acceleration of machine learning in the legal industry in recent years has been staggering. From chatbots that assist with research and drafting to predictive tools that identify patterns in legal data, the impact of these technologies is already being felt. A deeper philosophical question sits beneath that change: what happens to the concept of expertise when AI tools become increasingly adept at tasks that once served as proof of expertise? That may alter how expertise is valued within the profession and, in turn, how junior lawyers are perceived and compensated.

4. Example: Legal Research

As a junior lawyer at a digital law firm, I recently tested ChatGPT’s capabilities by asking it to draft a standard research memo. While the technology produced a draft quickly, the resulting document required substantial editing and refinement. Even so, it did save time in the research process by allowing us to focus more quickly on higher-level tasks. Legal technologies should still be approached with caution. They are not replacements for human expertise or judgment.

5. Review

From a technical standpoint, the memo provides a solid overview of the potential consequences of not possessing an Australian Financial Services Licence when engaging in crypto-asset transactions. But to improve its effectiveness, it would need more specific detail about the applicable legal penalties and business risks associated with operating without an AFSL. It would also need references and citations to support its claims and allow readers to verify the analysis. The text could be refined and condensed to improve concision and clarity.

One obvious critique is that the memo lacks a thorough treatment of the precise legal requirements governing crypto-assets in Australia, including the statutes, regulatory instruments, and licensing questions that actually matter in practice. It also does not adequately discuss the reputational damage that may follow from operating without a licence, nor does it provide guidance on how a company might obtain one.

6. Contracts and Drafting

Despite the many advances in AI technology, some sceptics still believe these systems merely repackage and plagiarise content from the web. While models can certainly process and generate vast amounts of text, the real measure of their sophistication is not simple regurgitation. It is the ability to produce original material tailored to a user’s needs.

In legal writing, for example, AI models such as ChatGPT are capable of generating contracts and clauses that are responsive to specific instructions. These clauses are not simply copied and pasted from existing sources. They are produced based on the user’s input and the model’s statistical understanding of legal language, structure, and terminology.

One criticism of the clause above is that it may be overly favourable to the contractor precisely because it was written by an AI system with no real-world appreciation for the nuance of contract law, commercial bargaining, or market norms. Without guidance from a human legal professional, the clause may lack the depth and precision typically required in real contracts.

The use of AI to generate contracts also raises concerns about accountability and liability. If the contract is defective or inadequate, it may be difficult to determine who is responsible: the AI, the person who fed instructions into the AI, or the party who agreed to the contract without fully understanding its terms. That uncertainty creates legal and practical risk for everyone involved.

7. The Catch

These are the first words I have typed myself, aside from the prompts. ChatGPT generated the outline for this concept, adapted the headings, and wrote nearly every paragraph above with only light grammatical editing. That includes the introduction, the critique, and the ethical concerns about its own output.

Perhaps more striking than the article itself is the red button waiting at the bottom of this section. I asked ChatGPT to write the code to hide text throughout an article, reveal that hidden text with a button press, and send the reader back to the top of the page. The model responded with the HTML, JavaScript, and implementation guidance required to make that experience work. I then asked it to make the prompt boxes feel more futuristic, to help brainstorm ideas, to tighten the article for readability, and to think about mobile responsiveness.

You might remain sceptical, and rightfully so. Any decent precedent bank inside a law firm would outperform much of the drafting above and would require less human intervention. But that is not really the point of this article. The point is efficiency. Tools like this move from curiosity to utility with unsettling speed, and even imperfect outputs can materially compress the time between question and first draft.

8. The Future Junior

Many recent Australian law graduates will already have spent time writing essays about disruption and the risks or opportunities created by rapidly evolving technology. Many are quick to defend the human touch that trained practitioners bring to the profession, particularly in areas such as family or criminal law. That instinct is understandable. The image of a cyborg judge deciding custody or imposing sentences is the sort of dystopian picture that makes people recoil.

However, the better question may not be whether technology can replace the lawyer. It may be what lawyers do when technology overtakes parts of their skillset.

This iteration of AI feels like the first step toward hyper-niche software-as-a-service products built specifically for law firms. Imagine a document management system fed every known case, piece of legislation, textbook, precedent, and resource in a tightly defined area of law. Its parameters adapt dynamically, creating new ways of interacting with the law from both a practitioner and a client perspective. A Residential Tenancies in Western Australia guidance kit could identify which clauses are strong, what a landlord should watch for, what a tenant should argue before a particular decision-maker, and how all of that should be framed in real time. A sole practitioner could move from matter to matter by adapting, checking, and editing machine-generated documents and correspondence from a single prompt.

For a nation already wrestling with a graduate glut in the legal industry, that future is not especially comforting. The advantages of this software do not favour the junior lawyer: it is available 24/7, produces work instantly, requires no training, no mentorship, no degree, and is almost certainly cheaper than paying a salary. Even in a middle-ground world where graduates work alongside AI, there is a risk that the junior lawyer loses the friction that once forced them to learn through drafting, revising, and thinking slowly about the work.

The AI itself points toward one of the biggest ethical problems: who do you sue if a contract goes south because the advice was false or incorrect? Existing legal systems are built around rights, responsibilities, and penalties that ultimately attach to a person. They do not sit neatly on top of a machine or an algorithm. That “responsible machine” problem will need to be addressed by governments sooner rather than later.

This article is not meant to be alarmist. It is meant to point out that the ground has already shifted. Knowledge workers across many industries are facing similar pressures, and the broader social consequences are still barely understood.

9. Prologue

As AI continues to advance, it is clear that the legal industry will not escape its disruptive effects. But the rise of AI does not have to spell the end of the profession. It can also be an opportunity to use new tools to improve performance, sharpen judgment, and drive positive change. And who knows, with AI beside us, the legal profession may even crack a smile every now and then. After all, they say laughter is the best medicine. Unless you are a lawyer, in which case it is probably a hefty bill.

What is the difference between a lawyer and a sentient AI?

One of them has a conscience.