The Same Chisel: What Japan’s Shokunin (職人) Can Teach Lawyers About AI
For thirteen centuries, Ise has rebuilt a structure that was not broken, at enormous cost, partly so that the hands that know how to build it never disappear.

The Same Chisel
Every carpenter in Japan can buy the same chisel. The same steel is sold to the weekend hobbyist and to the 宮大工(miya-dai-ku), the temple carpenters who build and repair Japan’s oldest wooden structures. In one pair of hands, the chisel makes a bookshelf. In another, it makes a joint that holds for a thousand years.
The chisel is ordinary. The hand is not.
AI just gave every lawyer the same chisel. The tools that summarize a record, draft a brief section, and surface the relevant cases in minutes are now available to both the solo practitioner and the 2,000-lawyer firm, to the first-year associate and the veteran trial lawyer alike. That raises a question the profession has never had to ask in quite this way: when everyone holds the same tool, what is a client actually paying for?
Japan’s craftsmen, 職人 (shoku-nin), have lived with versions of that question for centuries. Their history has something to say about ours.
Fourteen Centuries of Earned Reputation
In 578, Prince Shōtoku needed a Buddhist temple built. Japan had no carpenters trained in that kind of construction, so he brought them from the Korean kingdom of Baekje. One of them, Kongō Shigemitsu, stayed. The business he founded, Kongō Gumi, built Shitennō-ji temple in Osaka, and Kongō Gumi continued its work for more than 1,400 years. It is widely regarded as the oldest continuously operating company in the world.
That longevity was not the product of slick marketing. It came from work that could be inspected centuries later and still found to be sound.
The best-known modern example of Japan’s craftsmen is Tsunekazu Nishioka (1908–1995). Both his grandfather and his father had served as master carpenter, 棟梁 (tō-ryō), of Hōryū-ji temple, the seventh-century temple in Nara whose buildings are among the oldest surviving wooden structures in the world. Nishioka began learning the tools and the craft from his grandfather as a schoolboy and became an apprentice in his teens. He spent his career on Hōryū-ji, and on the reconstruction of Yakushi-ji temple’s main hall and west pagoda. He was stern enough to be nicknamed 鬼 (oni), “the devil.” At the same time, he was praised officially as a Person of Cultural Merit.
Nishioka understood the bond between a craftsman and his tools in a way worthy of reflection. “When you study someone’s tools, you come to know their skill,” he once said. The tool does not make the carpenter. But how a carpenter keeps and handles his tools reveals the carpenter.
That points to the deeper meaning of shokunin. The word is usually translated as “craftsman” or “artisan.” Toshio Odate, a master woodworker, asserted in his 1984 book on Japanese woodworking tools that the translation misses something. Skill alone was never the whole of it. Odate put it this way:
“The shokunin has a social obligation to work his best for the general welfare of the people. This obligation is both spiritual and material, in that no matter what it is, if society requires it, the shokunin’s responsibility is to fulfill the requirement.” Toshio Odate, Japanese Woodworking Tools: Their Tradition, Spirit and Use viii (1984).
The Preamble to the ABA Model Rules of Professional Conduct bears a striking resemblance: “A lawyer, as a member of the legal profession, is a representative of clients, an officer of the legal system and a public citizen having special responsibility for the quality of justice.” Model Rules of Pro. Conduct pmbl. ¶ 1 (Am. Bar Ass’n 2025).
Both texts define their profession by a duty that reaches past the paying customer: to the general welfare of the people in Odate’s telling, and to the quality of justice in the Model Rules.
So the question is not only what a client is paying for. Odate prompts a second one: what does society have a right to expect from the lawyer performing the work?
A System Built to Protect the Craft
No institution in Japan understood better than the administrators of Ise Jingū shrine that skill must be passed on deliberately.
Every twenty years, the principal shrine buildings at Ise are dismantled and rebuilt from scratch on an adjacent site, along with the sacred furnishings and garments. The ceremony is called 式年遷宮 (Shiki-nen Sen-gū). The tradition dates to the seventh century, and the first documented rebuilding took place in 690. The 62nd was completed in 2013. The next is set for 2033, and the preparatory rites have already begun.
The restorative cycle has its religious reasons. It also has a practical one that is often cited: twenty years is long enough to train the next generation, yet short enough that the masters who built the last shrine are still working to teach them. Stretch the interval to fifty years, and the knowledge could die with the people who hold it.
Consider what that means. For thirteen centuries, Ise has rebuilt a structure that was not broken, at enormous cost, partly so that the hands that know how to build it never disappear. The shrine is renewed. So is the craft.
The legal profession has never built a deliberate institution like Shikinen Sengū. It has had something cruder and less intentional: the work itself. Generations of lawyers learned to read a record because someone handed them a transcript and told them to read it. That accidental apprenticeship is exactly what AI now threatens to erase.
Same Tools, Different Hands
Every litigator now has tools that would have seemed like magic a few years ago. They summarize depositions. They chart claim language against prior art. They draft a passable section of a brief before the coffee is brewed.
When everyone has the tool, the tool stops being the difference. What remains is the hand.
In patent litigation, the hand is the ability to find the one fact in a 400-page record that actually decides the case. It is spotting the sentence in the prosecution history that the other side hoped no one would read. It is sitting with a bad fact long enough to know whether it is truly bad or merely uncomfortable. Patent cases can turn on a single word in a claim. I recently wrote about one that turned on a single carbon atom. No tool told anyone which detail was material. A trained eye did.
Nishioka’s point about tools applies here as well. A lawyer’s skill now shows partly in how the lawyer handles the new tools: what gets delegated and what does not, whether the lawyer can tell when a confident summary has smoothed over the one key detail, whether the citation that looks right gets checked.
A carpenter who does not know how his chisel fails will eventually ruin a joint. A lawyer who does not know how AI falters will eventually ruin a case.
The profession’s own rules reinforce the point. Since 2012, the comments to Model Rule 1.1 have made keeping abreast of “the benefits and risks associated with relevant technology” part of a lawyer’s duty of competence. Model Rules of Pro. Conduct r. 1.1 cmt. 8 (Am. Bar Ass’n 2025). In 2024, the ABA applied that duty to generative AI. Lawyers need not become experts in these tools, but they must reasonably understand their capabilities and limitations, and uncritical reliance on their output without appropriate independent verification can itself fall short of competence. ABA Comm. on Ethics & Pro. Resp., Formal Op. 512 (2024).
Nishioka would have recognized the intent of these rules. For a carpenter, knowing how the chisel fails is the mark of a craftsman. For a lawyer, it is an obligation.
The Apprenticeship Question
This is where the shokunin story becomes uncomfortable.
The traditional path to mastery in Japan was long, repetitive, and often deliberately inefficient. The apprentice learned by watching (見て盗む, mite-nusumu, “stealing with the eyes”) and spent years on preparatory work before touching anything important. Nishioka took only one apprentice, Mitsuo Ogawa, and at first, he turned Ogawa away. Ogawa was accepted only after training elsewhere, building household Buddhist altars. He then lived in the Nishioka household as a live-in apprentice.
Japanese arts describe this progression as 守破離 (shu-ha-ri): first follow the form exactly, then break from it, then leave it behind. The 守 “shu” stage is where the reps live. In law, it is also precisely the stage that AI does best for us: first-pass document review, the case summary, the chronology, the first draft.
So a common worry has taken hold. Experienced lawyers, the argument goes, will be well served by AI, because they built their judgment the slow way like Ogawa. Young lawyers will be at a disadvantage, because the work that used to train them is disappearing.
I am not convinced.
Judgment was never a matter of years. The “10,000-hour rule” that Malcolm Gladwell popularized in Outliers rests on research by the psychologist K. Anders Ericsson. Ericsson himself stressed that time alone is not the point. What builds expertise is deliberate practice: focused work at the edge of one’s ability, with feedback, repeated again and again. Some lawyers practice for thirty years and simply repeat their first year thirty times. Others become formidable in a fraction of that time, because every rep counted.
That path is still open to young lawyers. In some ways, it is wider. A young lawyer today can read more on-point opinions in a month than I could find and read in a year at the same stage. She can write her own analysis of a case, then have the tool argue the other side to find where her reasoning is thin. Feedback that once depended on a busy partner’s red pen is available on demand. And many younger lawyers are far more fluent with these tools than their seniors will ever be.
There is one condition, and it is not negotiable. The rep only counts if the lawyer’s own mind is flexed. Reading an AI summary of an opinion is not reading the opinion. Approving a machine-drafted argument is not building one. An apprentice who learns on a power tool still has to learn the grain of the wood. The tool does not excuse that, but used well, it speeds both the learning and the work. Used as a substitute, it produces a lawyer who has held the tools for a decade without ever learning the craft.
Perhaps he can build a bookshelf, but never an enduring shrine.
The Ones Who Adapted
The shokunin who survived did not do it by refusing new tools. They kept the core skill constant and changed almost everything around it.
Consider Tsubame, in Niigata. Its metalworking tradition began in the early 1600s, when a magistrate brought in smiths to teach local farmers to make traditional Japanese nails as a side trade. When Western nails displaced them in the Meiji era, Tsubame turned to hammered copperware and kiseru pipes. Later it turned to Western cutlery, where it came to hold more than 90 percent of Japan’s domestic market. Polishing skill developed on pipe stems moved to spoons. Eventually it moved to something no one in Edo could have imagined: the mirror-finished stainless back of the first iPod, released in 2001.
The iPod chapter carries a lesson of its own. The order was too large for any single shop, so the region’s polishers took it on together. Quality was a problem at first. Master polisher Mitsuo Tanaka later recalled that most of the craftsmen polished in their own idiosyncratic way (我流, ga-ryū). Results varied, and a large volume came back rejected. Tanaka developed a technique, taught it, and standardized it across the shops until the yield reached nearly 100 percent. The skill had always been there. What it lacked was a way to be shared.
Or consider Hosoo, a Nishijin textile house in Kyoto founded in 1688. Over roughly thirty years, the kimono industry shrank to about a tenth of its former size. The traditional Nishijin loom weaves fabric about 32 centimeters wide, the width of an obi sash. That is too narrow for walls or furniture without seams. When the architect Peter Marino wanted Nishijin techniques for his interiors, Hosoo developed a new loom that weaves about 150 centimeters wide, the international standard for textiles. Today, Hosoo fabrics line Dior and Chanel stores and luxury hotels around the world.
The weaving is still Nishijin weaving. The loom is new.
A Shikinen Sengū for the Legal Profession
Not every story ends well, and the lessons cut both ways. Even Kongō Gumi’s independence came to an end after fourteen centuries of business. In January 2006, financial trouble led to the sale of its assets to Takamatsu Construction Co., Ltd. Its temple carpenters work on, but the independent family firm did not survive.
Tsubame faces a different threat today. Its polishers and toolmakers are aging, successors are scarce, and the city has warned that closures could break the region’s supply chain. Tsubame’s response is instructive. It is developing automated polishing machines. At the same time, it runs a training center where certified masters teach the next generation. Machines and masters, together.
That is probably the right model for the legal profession too. AI will take over much of the work that once trained young lawyers, and in many cases it should. But the profession cannot assume that skill will pass itself on the way it did when the work was the apprenticeship. Ise rebuilds a shrine that is not broken so that the craft never dies. Law firms, senior lawyers, and young lawyers themselves will need something just as deliberate: reps chosen on purpose, done by a human, and checked by someone who knows the grain.
That is what society has a right to expect: not lawyers who refuse the new chisel, but lawyers who have earned the right to use it.





Comments