Powered by
MOMENTUM
MEDIA
The Australian legal profession has spent considerable time examining what artificial intelligence means for productivity, service delivery, and the economics of legal work. I think this question deserves equal attention: what happens when the work AI performs increasingly well is also the work we have traditionally relied upon to train junior lawyers? writes Jamie Ng.
Think back to the beginning of a legal career. There was the afternoon spent researching a difficult point, the first draft that came back covered in comments, and the precedent you initially thought was perfect until someone more experienced explained why it wasn’t. These tasks produced work for clients; but they were also part of an apprenticeship, and they required junior lawyers to wrestle with uncertainty, form a view, get things wrong, and gradually understand how an experienced practitioner approached the same problem.
AI can now compress many of those hours into minutes. Research can be assembled almost instantly, sophisticated first drafts can be produced before a junior has worked through the structure themselves, and large amounts of material can be summarised without the same process of reading, comparing, and deciding what matters.
That is an extraordinary productivity gain, but it also means Australian firms need to be much more conscious of separating efficiency in the work from development of the lawyer.
The pathway to judgement is changing
For generations, professional development was embedded in the production model of the firm. Junior lawyers learnt through doing because there was a commercial reason for them to do the work, and senior lawyers then corrected, challenged, and refined it because the client needed the work to be right.
As more of that first-pass production becomes AI-assisted, firms cannot simply assume the same learning will occur through a faster process and the skill we need to develop moves further towards interrogation. A junior lawyer has to understand what sits underneath an AI-generated answer: Which assumptions has the model made? What information has it weighted more heavily? What context may be absent, and is the apparently correct legal position actually useful for this particular client?
Those capabilities have always been part of good lawyering, but historically, they developed alongside technical training over time. We may now need to teach them earlier and with much greater intent.
The role of proximity
This also complicates the discussion about hybrid work because it would be easy to assume that more capable technology makes physical proximity progressively less important. In many respects, it does because lawyers can work across offices and jurisdictions more easily than at any point in the profession’s history – but professional development may move in the other direction.
If junior lawyers learn less through repetitive production, exposure to experienced lawyers becomes more important. Watching a partner explore a client’s commercial objectives, manage a disagreement, or explain why a technically correct answer does not work in practice gives a junior access to something no finished document can fully capture because these are the moments in which judgement becomes visible.
For firms thinking about the purpose of the office, it is worth considering that the future office may matter more as an environment in which experience, judgement, and professional behaviour are transferred between people.
Using AI without outsourcing judgement
This should not be taken as junior lawyers should be discouraged from using AI. Quite the opposite is true – they will need to become highly capable users of tools that will form part of ordinary legal practice. What matters is that capability around this technology is accompanied by the confidence to challenge it. Generative AI is extremely good at producing plausible answers, and often those answers will be correct. The more difficult professional problem arises when an answer is convincing but incomplete because a particular fact, client objective, or commercial consideration changes the appropriate course.
Lawyers have traditionally developed scepticism through researching competing positions, having their reasoning challenged, and learning that the conventional answer is not always sufficient. If AI removes some or even all of those steps, firms need to create the challenge elsewhere.
This could mean earlier client exposure, greater observation of senior practitioners, more discussion about why decisions were made, and more opportunities for junior lawyers to defend a view rather than simply produce an output. Senior lawyers may also need to become more deliberate about explaining their reasoning instead of only correcting the final work.
Reviewing improves the work. Coaching improves the professional. The firms that develop the strongest lawyers over the next decade will not be those that adopt AI fastest, but those that understand what the traditional work was teaching professionals and find better ways of preserving those lessons.
AI can accelerate the work considerably, but our responsibility is to make sure it also leaves a viable pathway through which today’s junior lawyers develop into tomorrow’s trusted advisers.
Jamie Ng is global clients and markets partner at Ashurst Perkins Coie.
