By researching a legal issue themselves, young lawyers learn how to locate and assess authorities. By preparing drafts, they learn how to construct legal arguments. And by having senior lawyers review and correct their work, they gradually develop the judgment required to practise independently. AI can now perform many of these tasks in a fraction of the time.

If AI performs the work through which junior lawyers traditionally learn, how will those lawyers acquire the experience needed to become competent senior lawyers? Judge Subramanian warned that firms need to think carefully about this consequence as they integrate AI into their practices. One possible approach, he suggested, would be for firms to require junior lawyers to prepare initial drafts without AI assistance before using the technology to refine or supplement their work. The objective would not be to ensure that lawyers continue developing the underlying skills that allow them to assess AI-generated work critically.

The risk extends to clients

A lawyer who has not developed sufficient experience may struggle to identify an AI-generated error. If that error makes its way into a court filing, legal advice or another important document, the consequences can fall directly on the client.

This creates a particular problem because AI-generated legal work can appear authoritative even when it is wrong. A fabricated case citation or inaccurate statement of law may not immediately look suspicious to an inexperienced lawyer. Detecting the problem requires the ability to question the output, check the underlying authority and understand whether the proposition actually follows from the law.

Those abilities come from legal training and experience. If junior lawyers increasingly rely on AI to perform the analytical work through which those abilities are developed, they may become less equipped to supervise the technology as they progress in their careers.

AI does not remove the lawyer's responsibility

The incident also reinforces a fundamental principle of professional responsibility: lawyers remain responsible for work submitted to a court even when artificial intelligence was involved in producing it. The use of an AI tool does not transfer professional responsibility to the technology. A lawyer must still verify authorities, quotations, legal propositions and factual assertions before relying on them.

This is particularly important because generative AI systems can produce plausible but incorrect legal material. The technology may assist with research or drafting, but the final responsibility remains with the lawyer. That responsibility becomes more difficult to discharge if lawyers are insufficiently trained to evaluate what AI produces.

The efficiency dilemma

AI can make legal work faster and potentially reduce costs for clients. Tasks that once required hours of junior-lawyer work can increasingly be completed in minutes. Clients may therefore have strong incentives to encourage firms to use AI. But the short-term efficiency gains may have longer-term consequences if they eliminate opportunities for junior lawyers to learn.

A law firm that uses AI to perform every first draft may become more efficient today. But it could also reduce the amount of substantive work through which its junior lawyers develop the expertise needed to handle complex matters tomorrow.

Source: Reuters