几代人以来，律所最宝贵的资产每晚都会走出大门。它是那位多年前记得拯救交易的条款的律师；是那位审判律师，他知道哪些论点对法官有效，哪些无效；是那位合伙人对助理说：“我们以前见过这种情况。”如今，律所正争相将这种专业知识转化为软件。在过去两年中采用了来自OpenAI、Harvey、微软、Legora等公司的工具后，大型律所正围绕着更有价值的东西构建系统：只有他们才知道的东西。
For generations, a law firm's most valuable asset walked out the door every night. It was the attorney who remembered the clause that saved a deal years ago. The trial lawyer who knew which arguments worked on a judge and which ones didn't. The partner who told an associate, "We've seen this before." Now law firms are racing to turn that expertise into software. After spending the past two years adopting tools from OpenAI, Harvey, Microsoft, Legora, and others, large law firms are building systems around something more valuable: what only they know.
他们的赌注是，融入他们交易历史、审判策略和客户偏好的人工智能将比街对面律所能购买的任何东西更有用。律所正在大笔投入，试图将他们的秘诀瓶装起来。今年早些时候，Kirkland & Ellis同意向Palantir支付5亿美元，为该所核心业务之一——为私募股权客户筹集资金提供咨询——构建技术。Latham & Watkins购买了自己的Nvidia服务器，以开始定制大型语言模型。
The bet is that artificial intelligence steeped in their deal histories, trial strategies, and client preferences will be more useful than whatever the firm across the street can buy. Law firms are spending big to bottle their secret sauce. Earlier this year, Kirkland & Ellis agreed to pay Palantir $500 million to build technology for one of the firm's core businesses: advising private equity clients on raising money. Latham & Watkins bought its own Nvidia servers to start customizing large language models.
在美国最大的人身伤害律所Morgan & Morgan，律师们已经开始使用一个内部人工智能模型，该模型经过训练，能够识别他们案件中的模式。一位年轻律师可能会错过：患者表现出脊椎附近神经受压的迹象，以后可能需要手术。该所表示，其软件可以标记这种可能性。“我们最优秀、经验最丰富的人几乎能立刻发现这一点，”Morgan & Morgan首席转型官Yath Ithayakumar说。
At Morgan & Morgan, America's largest personal injury law firm, lawyers have begun using an in-house artificial intelligence model trained to recognize patterns in their cases. A younger lawyer might miss that a patient showing signs of a pinched nerve near the spine could later require surgery. The firm says its software can flag that possibility. "Our best and experienced people will spot it almost immediately," said Yath Ithayakumar, Morgan & Morgan's chief transformation officer.
“但如果你是第一年律师，这是你的第一份工作，也许这对你来说并不那么明显。技术会带来巨大的差异。”这个故事是《AI数据掠夺》系列的一部分，深入探讨科技巨头如何日益转向工人、用户和物理世界，以获取数据来训练人工智能。该系列的其他故事请见此处。这就是Morgan & Morgan推动构建自身技术背后的承诺。
"But if you're a first-year attorney, this is your first job, maybe that's not as obvious to you. Technology makes a huge difference." This story is part of a series called The AI Data Grab, an inside look at how tech giants are increasingly turning to workers, users, and the physical world to get data to train AI.stories in this series here. That is the promise behind Morgan & Morgan's push to build its own technology.
这家公司表示将在未来十年投入10亿美元用于此项工作，它并不试图与OpenAI竞争。它希望捕捉其最优秀律师已有的知识。许多律师事务所已经在使用相同的高科技工具。Harvey表示，其用户覆盖全国按收入排名的前100大律师事务所的80%。Legora表示，其覆盖了这些事务所的大多数。当每个人都能购买相同的系统时，优势必须来自这些系统所掌握的知识，旧金山法律咨询集团LegalEng Consulting Group的首席执行官Mary O'Carroll表示，该公司是一家精品法律咨询公司。
The firm, which has said it will spend $1 billion on the effort over the next decade, isn't trying to compete with OpenAI. It wants to capture what its best lawyers already know. Many law firms already use the same high-tech tools. Harvey says it has users at 80% of the country's 100 largest law firms by revenue. Legora says it covers a majority of the same firms. When everyone can buy the same systems, the edge has to come from what those systems know, said Mary O'Carroll, chief executive of the LegalEng Consulting Group, a boutique legal consulting firm based in San Francisco.
如果一家律所“在时间长河中完成了所有私募股权交易，并垄断了市场”，那么其价值在于能够利用这段历史，使律师能够实际地说：“我口袋里装着5万笔交易的知识。”她这样说道。这有助于解释为什么Morgan & Morgan愿意走这么远，甚至自行开发技术。该律所拥有数十年的案卷，记录了其律师如何处理人身伤害案件、证据开示争端、和解及审判。
If a firm has "done every private equity deal for all of time and has a corner on the market," the prize is being able to use that history so lawyers can say, in effect, "I have the knowledge of 50,000 deals in my back pocket," she said. That helps explain why Morgan & Morgan is going so far as to build its own technology. The firm sits on decades of case files showing how its attorneys have handled injuries, discovery fights, settlements, and trials.
其历史不仅是工作记录，更是这些系统发挥作用所需的原始材料。“我们拥有的数据比其他任何人都多，”Ithayakumar说。“我们必须善加管理这些数据。”负责任的使用始于控制。Morgan & Morgan必须谨慎决定谁能访问其数据、他们被允许做什么，以及外部供应商是否能够保留或从中学习，Ithayakumar说道。而在该律所能够利用其数据跨案件发现模式之前，它必须先删除任何可能泄露当事人身份的细节。
Its history is not just a record of work done. It is the raw material these systems need to be useful. "We have more data than anybody else," Ithayakumar said. "We have to be good stewards of it." Responsible use starts with control. Morgan & Morgan has to be careful about who gets access to its data, what they're allowed to do with it, and whether outside vendors can retain or learn from it, Ithayakumar said. And before the firm can use its data to spot patterns across cases, it has to strip away any details that could identify the people behind them.
这种对控制的需求为科技公司开辟了新的方向。Harvey希望为律师事务所做Palantir为Kirkland & Ellis所做的事情。今年，Harvey采用了一个开放权重模型，并对其进行了进一步训练以适用于法律工作。最终，它希望帮助律所用类似的方式处理自身事务。LexisNexis和Thomson Reuters（Westlaw的所有者）也看到了类似的机遇。
That need for control has opened a new lane for tech companies. Harvey wants to do for law firms what Palantir is doing for Kirkland & Ellis. This year, Harvey took an open-weight model and trained it further for legal work. Eventually, it wants to help firms do something similar with their own matters. LexisNexis and Thomson Reuters, owner of Westlaw, see a similar opening.
两家公司都通过收集案例、法规和条例，然后在此基础上加入他们自己的编辑分析，从而建立了庞大的业务。随后，他们进军律师用于起草、研究和审阅文件的软件领域。如今，他们正逐步接近这些模型本身。汤森路透正向客户提供对汤森的访问权限，该模型是在其150年内容的基础上训练得出的。LexisNexis表示，它也在朝着类似的未来努力：模型将同时基于其法律语料库和律所自身的数据。
Both built massive businesses by collecting cases, statutes, and regulations, then layering them with their own editorial analysis. Next, they pushed into the software lawyers use to draft, research, and review documents. Now they are moving closer to the models themselves. Thomson Reuters is offering customers access to Thomson, a model trained on 150 years of its own content. LexisNexis says it's aiming for a similar future: models grounded in both its legal corpus and a firm's own data.
律所正是提出了这一需求，LexisNexis全球法律业务首席执行官肖恩·菲茨帕特里克如是说。“所有模型都能以类似的方式完成任务，”他说。使它们有所区别的“是这些模型所拥有的内容和上下文。”尽管这些定制系统正吸引大量资金投入，但目前还为时过早，无法判断胜出者将是拥有最多数据的律所，还是帮助他们使用数据的软件公司。
Law firms asked for it, said Sean Fitzpatrick, chief executive of global legal at LexisNexis. "All of the models can do things in a similar way," he said. What sets them apart "is the content and the context that those models have." For all the money flowing into these custom systems, it's too soon to tell whether the winners will be the firms with the most data or the software companies that help them use it.
但所有人似乎都同意发展方向。法律技术领域的下一场较量，不再是关于谁拥有最智能的软件，而是关于谁能掌控软件所掌握的知识。
What everyone seems to agree on is the direction of travel. The next fight in legal technology is not over who has the smartest software. It's over who gets to own what the software knows.