Series Index
Part 1 | Attorney Hong Seung-kwon: “AI Arrives in the Legal Profession”
Part 2 | How Should Lawyers Use AI?
Part 3 | Dividing Roles Between AI and Lawyers
Part 4 | AI Legal Advice Isn’t Enough—Here’s Why Lawyers Still Matter
By Hong Seung-kwon
Principal Attorney, PAN Law Firm (skh@law-pan.com)
Today, I posed another question to the AI system I frequently use. The structure of the case was simple, but the legal outcome could hinge on subtle distinctions.
“A creditor received a garnishment and collection order over a claim the debtor holds against a third party and subsequently collected payment from that third party. However, the creditor did not file a collection report and is now demanding payment again from the debtor based on the original claim. Since the failure to report the collection means the creditor’s claim against the debtor has not technically been extinguished, does the debtor have any legal grounds to refuse the creditor’s demand?”
The conclusion aligns with my initial assessment. However, along the way, the AI made an incorrect judgment in its reasoning process.
So I asked several follow-up questions. Through this iterative exchange, I was able to correct the errors and ultimately confirm that my original conclusion was indeed correct.
This experience illustrates both the limitations and the potential of today’s AI. The final answer was accurate, but the reasoning contained flaws, and it took several rounds of clarification to arrive at a reliable conclusion.
Part of the AI’s error may have stemmed from the fact that my question was not sufficiently precise. Yet asking precise questions is not easy.
We are often busy, and crafting carefully structured sentences requires significant mental energy. Clients are no different.
In legal consultations, clients rarely present perfectly organized questions. Their accounts are often emotional, legally imprecise, and sometimes omit facts that may be unfavorable—whether intentionally or not.
Lawyers work within this imperfect flow of information. They identify the core issues and uncover the essence of a case through targeted, strategic questioning.
In this sense, AI and lawyers can be ideal partners. AI can rapidly review vast volumes of case law and statutes, identify analogous precedents, and propose multiple legal interpretations.
Lawyers, in turn, develop strategies tailored to a client’s specific circumstances based on that information. They also correct AI’s mistakes and determine how conclusions should ultimately be framed and applied.
Today’s experience was a clear example of such collaboration. I detected errors in the AI’s initial answer, refined the inquiry through follow-up questions, and reached the correct final conclusion.
Of course, AI is still far from perfectly understanding ambiguous questions, let alone providing consistently accurate answers. For AI to become a truly effective legal tool, it must learn to discern what a questioner genuinely intends to ask—even when the question itself is unclear.
It must also be able to provide the precise information needed. This remains one of AI’s core challenges and a central issue for its application in the legal field.
Just as AI must improve its ability to interpret imperfect questions, lawyers must develop the skills to collaborate effectively with AI. Knowing what to ask AI in order to obtain useful answers is itself becoming a new professional competency.
Equally important is knowing which parts of AI’s answers require verification and where AI may overlook critical nuances. In the future, the ability to design processes of “verification” will be as important as understanding legal doctrine itself.
Ultimately, the path forward for AI and lawyers lies in genuine collaboration—recognizing each other’s strengths and compensating for each other’s weaknesses. AI must advance toward a deeper understanding of human context.
Lawyers, meanwhile, must adapt their ability to utilize AI effectively. When these two trajectories converge, legal services can become more accurate, efficient, and accessible.
If AI eventually progresses beyond merely knowing statutes and precedents—if it can fully understand a client’s situation and identify the best solution between theory and real-world constraints—could it replace lawyers?
Honestly, perhaps it could. As I have written before, a lawyer’s most important role is not simply knowing which legal arguments are possible, but deciding whether those arguments should actually be made.
If AI reaches the point where it can replace lawyers, society itself will already be fundamentally different from what it is today. And when that moment comes, we may need to ask not only what role lawyers will play—but what role we, as humans, should take on next.
