How Will Young Attorneys Develop in the Age of AI?

Hello, I'm Noriaki Asato, Representative Attorney at LegalAgent.

At LegalAgent, attorneys review drafts and research results prepared by generative AI and put them to use in handling matters.

Legal Agent, a Law Firm for the Generative AI Era

LegalAgent uses generative AI and commits to responding within one business day as a rule. Deadlines for deliverables are adjusted individually according to the content and volume of each matter. Recently, as I think about the use of generative AI at law firms, I have had more occasions to confront the question of how young attorneys will develop.

Replacing First-Stage Work and Changes in Training Opportunities

First, to be frank, I think generative AI will greatly change the work that many associate attorneys have handled.

I felt this when I built the Word add-in too, but working with Codex, the "associates are unnecessary" argument is becoming more and more realistic.
That said, I myself was once an "associate attorney" trained by senior colleagues and clients, and in terms of "the sustainable growth of the Japanese legal profession" as well, ...
— Noriaki Asato (Asato Noriaki) @Legal Agent (@NoriakiAsato)
May 2, 2026

The change I mean here is not that attorney qualifications themselves or the final judgment of a professional will become unnecessary. It means that, in corporate legal practice, a substantial part of the work that associate attorneys, mainly in their first to third years, have spent long hours on will be replaced by generative AI. Of course, this does not mean that all of young attorneys' work will disappear, nor that AI will immediately guarantee deliverables of the same quality in every case.

Young attorneys who have just joined a law firm are not in a position to decide the overall strategy for matters from the start. They enter practice by reading materials, reviewing contracts, and preparing first drafts and research memos. This work was the firm's productive activity and, at the same time, an important opportunity for young attorneys to learn the basics. Yet generative AI shows high processing capability in exactly this area. With AI, one can read large volumes of material, compare differences between clauses, search past templates and similar matters, and prepare draft text and candidate issues.

As a result, the move to ask AI to handle first-stage work that used to be entrusted to young attorneys is spreading. If a first draft of a certain standard can be obtained in a short time, clients find it harder to see value in spending a long time on the first-stage work itself. For the firm, too, it becomes difficult to justify deliberately assigning people to work that AI can handle and billing for the time.

This change also affects the pyramid-shaped organizational structure in which partners oversee the whole and many associates support the volume of work. I think the idea of handling volume by adding more young attorneys will weaken, and things will shift to a form in which a small number of attorneys use AI to move first-stage work forward quickly and attorneys make the final legal judgments.

Working carefully remains important in itself, but if a person simply spends hours redoing work at the same level of detail as a first draft AI produced in a short time, that does not add enough value. Until now, young attorneys developed a feel for practice through the patient process of rereading contracts many times, struggling with research, searching past matters and receiving redlines from senior colleagues. If the first-stage process is replaced by AI, there is a concern that such natural training opportunities will decrease.

The Turnover of Matters and the Accumulation of Experience

That said, I am not simply pessimistic about this change. AI lowers the value of time-consuming work itself, but it raises the value of the ability to check deliverables created by AI, question them, revise them and turn them into concrete judgments. This ability, too, grows through daily practical experience.

In corporate legal work, the quality of an attorney's judgment depends greatly on experience. Even with the same service agreement, an attorney who has handled only a few and one who has seen a hundred or a thousand differ in how deeply they can read the contract. Judgments such as whether to set a cap on liability for damages or whether to exclude breaches of confidentiality obligations from it are supported not only by knowledge of the provisions but also by experience of what disputes and practical problems arose in similar past matters.

Using generative AI speeds up summarizing contracts, comparing clauses and drafting comments, and increases the volume of material that can be reviewed in due diligence. There used to be a limit to the number of matters and contracts a single young attorney could handle in a year, but shortening the time spent on preparatory steps creates room to increase young attorneys' opportunities to engage with actual practice.

For example, it is precisely by reviewing dozens of NDAs that one comes to understand the differences in how clauses on use for other purposes are worded. And by looking at the same investment agreement repeatedly, one comes to see where the clauses that truly weigh heavily on founders lie. With AI handling the groundwork, young attorneys can work through each matter faster, compare a variety of deliverables and receive feedback from senior colleagues more often. Simply introducing AI does not automatically improve practical skills, but positioning AI as a tool that raises the turnover of experience opens a path to faster growth.

Verifying Output and the Burden of Review

When bringing AI into work, one must pay attention to the difficulty of rapidly checking the large volume of deliverables AI produces and continuing to make judgments. AI puts text into proper form and returns output that, at first glance, appears to cover every issue. But in legal practice, being well formatted is not enough.

It is essential to check whether the statutes and official guidelines on which the output is based are free of errors, whether the underlying facts have been misunderstood, and whether it fits the client's business objectives and position. It is also important to properly separate comments that convey the identified risks to the counterparty as they are from issues that should be kept as internal notes. Whether there are issues that should deliberately not be raised is also considered in light of the background of the transaction, the relationship with the counterparty and risk tolerance.

This verification work carries a heavy mental and physical burden. I myself sometimes feel as though my mind cannot keep up when rapidly checking the draft contracts and research results that AI presents one after another. As young attorneys face this burden, it is essential that they have senior colleagues and peers with whom they can frankly discuss questions rather than carrying everything alone.

The ability not to accept AI output uncritically, to obtain support from primary sources and to raise questions while comparing with past cases is honed precisely in practical experience. Rather than keeping AI at arm's length, it is effective to build up training in actually reading AI output, revising it repeatedly and receiving concrete feedback from senior colleagues.

Deliberate Training Design and Firm Support

At law firms, there are many opportunities for young attorneys to learn through handling day-to-day matters. On the other hand, experience can vary depending on the matters assigned and the opportunities to receive guidance. But in an age when the spread of AI shortens work processes, it is important not to leave training to chance and to deliberately put it in place as a system.

Specifically, we are moving toward a practice of recording from what perspectives young attorneys reviewed deliverables generated by AI, which issues they revised, and why. If the reasons why senior attorneys made revisions, and with what intent, accumulate in matter folders and review histories, young attorneys can not only compare AI output with the finished version but also learn in depth the thought process that leads to a judgment.

If there is an environment in which young attorneys engage with many matters from an early stage, check AI output and keep learning from the reasons behind senior colleagues' judgments, they should be able to acquire practical skills at a faster pace than before. At LegalAgent, too, while deepening collaboration between generative AI and attorneys, we will continue to value creating an environment in which the next generation of practitioners can grow steadily.

Our firm handles contract review, fundraising, M&A and more through collaboration between generative AI and attorneys well versed in corporate law. If you have any legal concerns, I believe there are many situations in which we can help, so please feel free to contact us.

Legal Agent, a Law Firm for the Generative AI Era