AI for Lawyers
AI for Lawyers

AI for Lawyers

Table of Contents

In M&A, it can be change of control, assignment, exclusivity, and the like. (Location 386)

Tags: ddas

Note: .ddas

In real estate, it might be base and additional rent, subletting, or maintenance responsibilities. (Location 387)

Tags: realestate, ddas

Note: .ddas .realestate

PART I The Point: AI in law is here to stay. It's time to take advantage

CHAPTER 1 How Lawyers Learned to Stop Worrying and Love AI

The questions and decision trees in these systems must be handcrafted by human experts, generally falling into the “rule based” or “reasoning” subfield of AI. Expert systems (Location 548)

Tags: expertsystems

Note: .expertsystems

In the legal world, AI is being used for - contract drafting - negotiation - review - litigation document review and analysis - predicting case outcomes - suggesting courses of action - organizing legal research - time keeping It is opening up possibilities never before imagined and allowing lawyers to spend more time on law and less time on repetitive activities. AI is partnering with lawyers, rather than replacing them. (Location 560)

Tags: ai

Note: .ai

Recurring issues we've found lawyers raise regarding AI are: “How can I trust AI software?” “What if our associates use the tech to ‘cheat'?” “How are new lawyers going to grow into great lawyers with technology doing their work?” “Will using AI software impact (i) my duty to keep client information confidential, or (ii) lawyer–client privilege?” “Do I have to invest a lot of time in training AI to get value out of it?” “If AI makes lawyers way more efficient, will we need fewer lawyers?” “How does being more efficient work out for me if I bill hourly?” “How do I justify the extra expense of the software?” (Location 595)

Tags: ai

Note: .ai

You could do a keyword search (ctrl-f) for relevant words such as “assignment” or a phrase like “most favored customer.” The problem is that important concepts like “change of control” or “exclusivity” are frequently phrased in nonstandard ways, which makes keyword searching risky. Worse, many contracts for review come in the form of poor quality scans. (Location 668)

Tags: ddas

Note: .ddas ctrl f is less useful when there are multiple ways to spell and documents are low quality

CHAPTER 2 #DoMoreLaw: How Doing Work More Efficiently Can Create More Legal Work, Not Less

Jevons paradox. Economist William Stanley Jevons saw how, during the First Industrial Revolution, technology was making coal usage more efficient over time. Surprisingly, this led to more coal usage. Efficient technology made coal effectively cheaper to run, leading to more possible uses, counterintuitively increasing overall coal consumption. (Location 968)

Tags: efficiency, hc45, jevonsparadox

Note: .jevonsparadox as things become more efficient demand may rise

a law firm's biggest competitor is often its own clients, as those clients find ways to get better value by solving legal problems themselves with the help of their own staff and technology. (Location 1006)

Tags: lawfirms

Note: .lawfirms clliennts are frequently looking to take woork inhouse to cut costs

thanks to AI, total diligence is now possible. Rather than a 10% sample, you can now review 25%, 50%, or even 100% of contracts in question, in a manageable amount of time, for an acceptable (though not necessarily low) amount of money. (Location 1043)

Tags: ddas

Note: .ddas

Typically, there are two types of company buyers: financial and strategic. Financial buyers (like private equity firms) are concerned with buying and reselling businesses at a gain. Since they tend to buy businesses and run them as is (in an isolated legal entity), financial buyers are less likely to have issues with a target company's contracts. However, they can still benefit from a faster and deeper review. A faster, light contract review early in the evaluation period can help financial buyers determine which deals to lean into. Also, findings from more thorough diligence can allow financial buyers to more accurately set a fair price for the asset. For example, if an exclusivity clause limits a target's scope of operations, its value may be impaired. (Location 1052)

Tags: pe, ma

Note: .ma .pe

Strategic buyers, on the other hand, add companies they buy into their already-running businesses. This significantly raises the stakes on contract review. Contracts of an acquired entity are equally binding as those the company enters into in the ordinary course of business. Many companies put a lot of effort into ensuring that new contracts they enter into are properly approved. They should be equally careful about contracts they acquire in M&A. Sometimes, really bad things lurk in contracts, even seemingly inconsequential ones. Exclusivity. Non compete. Most favored customer pricing, indemnifications, uncapped liabilities, data transfer restrictions, and other clauses you might never find unless you dig down and look closely. These risks can compound when brought under a large acquirer's significant corporate umbrella. Imagine an emerging beverage company. If things go wrong as it grows, this could wipe out the company, but the company might not be that big so losses are naturally limited. The company is, to a certain extent, “judgment proof.” If Coca-Cola or PepsiCo buy them, all of a sudden there is a whole lot more to lose. (Location 1057)

Tags: ma

Note: .ma

Lawyers can add value by using AI to increase the number of contracts they review in transactions. Some clients might be happy to get a lower diligence bill thanks to faster AI-enhanced contract review. But many should be very interested in getting twice the diligence for the same price they paid the last time they did a deal, or three times for 30% (or 50%) more money. (Location 1067)

Tags: ai, ddas

Note: .ddas .ai

While Noah would like to think that they hosted him and his ex-colleagues just to catch up, he suspects the real purpose was to drive deals and litigation projects from alums who had moved in-house. (Location 1136)

Tags: alumni

Note: .alumni

In the course of the review, it became clear that, worldwide, there were far more legal documents on file than initially expected. In sum, the 1 million that were initially expected to be reviewed was only 25% of the 4 million active documents. However, the project team kept their heads down and got it done. (Location 1208)

Note: The scope of contracts was four times larger than expected

There are well over a million lawyers in the United States. More than 150,000 barristers and solicitors in England and Wales. More than 130,000 in Canada. Over 160,000 in Germany. Some 800,000 in Brazil. (Location 1240)

Tags: lawyers

Note: .lawyers

CHAPTER 3 The Small Law Mindset: Bringing Biglaw Capabilities to Smaller Firms

CHAPTER 4 AI: A Modern Job Creator: The New Legal Mindsets, Skill Sets, and Jobs That Are in Demand

In the legal field, the Jevons paradox means that as AI and other technologies make the delivery of legal services more efficient, that efficiency will drive more legal work—in the same way that more efficient refrigeration enabled an explosion of economic activity around frozen and refrigerated foods, not to mention more building of refrigerators themselves. (Location 1488)

Tags: jevonsparadox

Note: .jevonsparadox

The primary job in the legal services industry has been the licensed lawyer or attorney. These professionals have typically been engaged in roles in the following broad categories: Private practice attorneys in law firms, including sole practitioners. Corporate lawyers, on in-house legal teams in corporations. Lawyers engaged in representing parties in the justice system: prosecutors, public defenders. Lawyers engaged in public policy: advisors to legislatures, regional and local governments, government agencies, public officials, and transnational and nongovernmental organizations. Judges, arbitrators, and other roles involved in adjudicating disputes. (Location 1500)

Tags: lawyers

Note: .lawyers

Innovation and strategy roles. As legal organizations have turned to technology and new business models to support their delivery of legal services, new jobs have emerged for people capable of driving and supporting innovation, providing change management, and staking out a strategic direction. (Location 1528)

Tags: changemanagement, innovation

Note: .innovation .changemanagement

One result of this disaggregation of work is the increasing need for legal project managers who can bring all the moving pieces and parties together by using the three key components: people, process, and technology. (Location 1579)

Corporate legal departments are contending with a multitude of interrelated and overlapping market forces: The volume of work continues to increase exponentially as organizations deal with an explosion in data, rapid changes in regulations, and an increasingly interconnected global marketplace. In-house teams lack the bandwidth to keep pace. Additionally, most are trained in the practice of law, not the business of delivering legal services. All the while, the corporate legal department is not only being asked to do more with less but also to go a step further and partner with business to drive value. (Location 1721)

This problem-solving approach focuses on looking for patterns. The training shifts your mindset. If something looks unique, step back until it looks like something you have seen before. The patterns and the cycles eventually connect and reveal how things interact as a system. Most complex problems are systems problems that require solution approaches that address the bigger picture. (Location 1818)

A frequent innovation-focused conversation with attorneys starts off discussing how their work might be done differently using common patterns or tools and often produces, “Oh, but you don't understand. What I do here is different every time. It is not reducible to any of these more generalized models.” They will then tell you why a pattern approach is appropriate for some other legal work outside their practice, which they regard as too complex. The building blocks of most legal practices reveal repetition with patterns. This does not mean the work is not complex. It means the experts can benefit from assistance that eases their process burdens. (Location 1827)

We optimize work by adding process and technology to the parts that make sense. Many legal professionals presume process and automation must cover every scenario they might see. They then consider how challenging that would be, and dismiss the idea that there might be substantial value in focusing on the 80% of the work that reflects known patterns. This automatable portion is usually the work they enjoy the least, because it requires less critical thinking. But identifying opportunities to transform their practice requires reflecting on the work as a system and recognizing the parts that have patterns that are materially similar, despite the deviations. This is a teachable skill. (Location 1832)

Tags: automate

Note: .automate automate 80%

CHAPTER 5 Amplifying Legal Expertise: From One-to-One Advisory to Scalable, One-to-Many Service Delivery

For a law firm or corporate legal department, the use of technology to gather, maintain, and train lawyers in how they do things as an organization can help establish the organization as a unique entity. This also means that attorneys can't just go to another firm and necessarily be as successful. (Location 2041)

CHAPTER 6 The Ethics of Lawyers Using—and Not Using—AI

PART II The Proof: AI technologies are delivering results today

CHAPTER 7 eDiscovery : What It Is and How AI Is Continuing to Transform How It Works for You

lawyers have at their disposal three forms of evidence. The first is testimonial evidence, which is what we typically see on TV. This kind of evidence involves a witness telling their story or version of events, and a lawyer cross-examining that witness to poke holes in the story, to test it for reliability. A second form of evidence is demonstrative, which consists of actual evidence (such as a bloody glove from a murder trial) or an illustration of evidence (such as a blueprint of a building, or an illustration of a how a machine works). Demonstrative evidence is things brought into a courtroom, or places that a jury is taken to get a better sense of what might have happened. A form of demonstrative evidence that stands on its own is the third kind of evidence and that's documentary evidence, which can be understood as testimony in written form. Contracts, letters, purchase orders, or anything reduced to writing might be documentary evidence. (Location 2429)

Tags: evidence, litigation

Note: .litigation .evidence testimony,demonstrative,documentary

eDiscovery really consists of several distinct but related steps, and these have been cataloged in something called the Electronic Discovery Reference Model (or EDRM). (Location 2506)

Tags: edrm, ediscovery

Note: .ediscovery .edrm

eDiscovery technology can not only be used to train a system to retrieve text within documents likely to be responsive, but now have other functionality, such as: the ability to create timelines of who communicated with whom; to pull out and identify proper or place names and put them (and their associated documents) on a map; the ability to annotate, redact, and highlight documents; and the ability to seamlessly move between various document types (i.e., to review a tweet, an email, a text message, and an Excel file all in the same viewer and system). Systems are also emerging to detect “sentiment” (positive or negative feelings being communicated) within email and message communications, to enable human reviewers to focus on communications of emotional interest as categorized by the computer system. (Location 2648)

Tags: ediscovery

Note: .ediscovery

CHAPTER 8 AI in Legal Research: How AI Is Providing Everyone Access to Information and Leveling the Playing Field for Firms of All Sizes

Fundamentally, legal research is an exercise in determining what the law is—information that is then used to advise a client, craft contract language, or persuade a court that it should rule in favor of one party rather than another. (Location 2702)

conceptual searching, is equally important. Conceptual searching means that the search program “understands” what you are looking for and finds relevant material, even if the language you use to search is dramatically different from the language in the search result. This means that an attorney can search using their own language without fear that they will miss out on a relevant precedent that happens to articulate the same concept differently. For example, a search about “not earning a diploma” might return results regarding “failing to graduate” or “not completing academic training.” (Location 2817)

Tags: search

Note: .search Conceptual search is like seaching for woords with similar meaning too

Casetext is to our knowledge the only company that has released a full-fledged legal research engine based on the BERT approach (released as a tool called “Parallel Search”). (Location 2837)

Tags: search

Note: .search

One thing we found fascinating was the “features” (words/phrases) our NLP systems found predictive of a case being overruled. For example, the word today was suggestive as judges, perhaps in a bit of fanfare, would often announce that they “today” render an earlier holding obsolete. There were many more examples where the machine taught the lawyers about patterns we never would have guessed. (Location 2857)

Tags: proxy, judgements

Note: .judgements .proxy certain words,such as "today", are often used when a precedent is being overruled

Litigation Analytics. Allows lawyers to predict outcomes by mining past court dockets for insights about judges, parties, courts, and areas of law. It leverages visualization techniques and uses AI to extract and normalize data from thousands of courts. (Location 2894)

Tags: westlaw, litigation

Note: .litigation .westlaw

electric light the most efficient policeman.” (Location 3063)

Tags: police, light

Note: .light .police light is the best deterant for crime

clean, legally accurate data helps catalyze not only initial adjustments, but ongoing optimization over time: both for actors within those systems (client outcomes) and systems as a whole. In short, legal “history” helps us create better outcomes and avoid the mistakes of the past. (Location 3083)

Tags: data

Note: .data

Blue J Legal provides a predictive analytics platform targeted to the tax law space. (Location 3133)

Tags: taxi, taxlaw

Note: .taxlaw .taxi

Contracts are perhaps the most important documents in the world in terms of their impact and governance on all of our fundamental business relationships. (Location 3224)

Tags: blog post, contracts

Note: .contracts

Performance on known and unknown documents is a big dividing line in automated contract review software systems. If all you seek is data from known documents, pretty much any system should be able to meet your technical needs, and you can make a decision based on other user experience factors. If, however, you need to review unfamiliar documents, you will need a system that can accurately extract data from unfamiliar documents, and not every system will perform at the same level. (Location 3371)

Tags: ai, kira

Note: .kira .ai

What we found was that with the AI assist, the team was spending 20–60% less time on the same tasks that they had done manually prior to using AI. (The median time-savings was 50%.) Of course, contract review doesn't impact every task in due diligence, so this is not to say that clients’ due diligence bills went down by 50%, but for every task that utilized AI in such a capacity, the median time saved was 50%. (Location 3410)

Tags: ddas

Note: .ddas

We see AI driving three big changes in the contract analysis space: An increase in the size and scope of contract review for law firms and enterprises. ... We will see more AI contract applications. We are seeing drafting and negotiation tools with embedded playbooks, but no doubt we will start to see other, more interesting tools that we can't even imagine. (Location 3482)

Tags: ai, kira

Note: .kira .ai

AI in law, across all its many forms, is fundamentally about one intellectual process—inference: if A, then B. Laws and regulations are written as inferences. Law schools train lawyers in legal inference. In practice, lawyers make inferences all day. That's what, in a sense, clients engage lawyers to do. They build if-then statements in their minds. If there is some fact or pattern of facts that we can observe with some degree of accuracy and certainty, then we can conclude some other things. (Location 3508)

Tags: law, expertsystems

Note: .expertsystems .law

In AI in the law field, there are five principal avenues to value: Legal research Electronic discovery Contract analysis—pre-signature and post-signature Predictive analytics—court processes and outcomes; pricing and legal operations; government processes Expertise automation (Location 3548)

Tags: expertsystems, ai

Note: .ai .expertsystems

A firm with a large regulatory practice might spend a lot of time with clients answering routine questions about the application of regulatory regimes to their day-to-day business practices. That firm might benefit from creating rules-based expert systems that allow clients a certain level of self-service for those kinds of questions. (Location 3741)

Tags: expertsystems

Note: .expertsystems