Key References
https://www.brookings.edu/blog/techtank/2023/03/20/how-ai-will-revolutionize-the-practice-of-law/
- Law firms that effectively leverage emerging AI technologies will be able to offer services at lower cost, higher efficiency, and with higher odds of favorable outcomes in litigation.
- Consider one of the most time-consuming tasks in litigation: extracting structure, meaning, and salient information from an enormous set of documents produced during discovery
- Or consider the drafting of motions to file with a court. AI can be used to very quickly produce initial drafts, citing the relevant case law, advancing arguments, and rebutting (as well as anticipating) arguments advanced by opposing counsel.
- more efficient for attorneys to draft documents requiring a high degree of customization….. include contracts, the many different types of documents that get filed with a court in litigation, responses to interrogatories, summaries for clients of recent developments in an ongoing legal matter, visual aids for use in trial, and pitches aimed at landing new clients.
- AI could also be used during a trial to analyze a trial transcript in real time and provide input to attorneys that can help them choose which questions to ask witnesses.
- LegalTech:
- Casetext - CoCounsel (technology from OpenAI)
- https://www.youtube.com/watch?v=ZKLkmdK_Odw
- “Can you research what courts in this jurisdiction have done in cases presenting similar fact patterns to the case we are working on?”
- Skills needed:
- knowing how to choose the right AI tool for a particular task, knowing how to construct the right queries and evaluate the relevance, quality, and accuracy of the responses
- use of AI tools is done with appropriate attention to protecting confidentiality
- Impact:
- AI also has the potential to dramatically broaden access to legal services, which are prohibitively expensive for many individuals and small businesses
- AI will make it much less costly to initiate and pursue litigation. For instance, it is now possible with one click to automatically generate a 1000-word lawsuit against robocallers.
- flood court systems in multiple jurisdictions with frivolous AI-written lawsuits
- AI Cannot replace:
- convincing presentation to a jury
- can’t fully weigh the factors that go into the many strategic decisions
- human element of relationships with clients
- leadership role in motivating a team of attorneys to produce their best work
https://www.nytimes.com/2023/04/10/technology/ai-is-coming-for-lawyers-again.html
- “That is really, really powerful,” said Robert Plotkin, an intellectual property lawyer in Cambridge, Mass. “My work and my career has been mostly writing text.”
- One new study , by researchers at Princeton University, the University of Pennsylvania and New York University, concluded that the industry most exposed to the new A.I. was “legal services.”
- “The reality is A.I. has not disrupted the legal industry,” said Ben Allgrove, a partner at the firm and its chief innovation officer. The rapid progress in large language models — the technology engine for ChatGPT — is a significant advance, Mr. Allgrove said. Reading, analyzing and summarizing, he said, are fundamental legal skills. “At its best, the technology seems like a very smart paralegal, and it will improve,” he said. The impact, Mr. Allgrove said, will be to force everyone in the profession, from paralegals to $1,000-an-hour partners, to move up the skills ladder to stay ahead of the technology.
- And the software models’ tendency to make up things confidently is alarming — and an invitation to malpractice suits — in a profession that hinges on finding and weighing facts. To help address those concerns, law firms often use software that runs on top of something like ChatGPT and is fine-tuned for legal work. The tailored software has been developed by legal tech start-ups like Casetext and Harvey.
- Lawyers at big firms have seen significant time savings for certain jobs and view the technology as a tool to make teams of lawyers and paralegals more productive.
- Mr. Washington used the software in a suit against the City of Flint claiming that residents were overcharged on water and sewer rates and service fees. He loaded more than 400 pages of documents, and the software quickly reviewed them and wrote a summary that pointed him to an important gap in the defense’s case.The program did in a few minutes what would have taken him several hours, he said.
- LegalTech:
- Harvey:
- Business Model:
- Higher productivity means fewer billable hours, yet hourly billing remains the dominant business model in legal work. A.I. should increase the pressure from corporate clients to pay law firms for work done rather than time spent.
- Comments:
- I recently tried out CoCounsel from Casetext. It's a great tool and extremely useful for certain tasks (quick research answers!) but it currently has some real structural limitations inherent to how the underlying GPT4 model works when it comes to document analysis. Take analyzing a deposition transcript, often hundreds of pages. GPT4's context window is limited to either 8k or 32k tokens, depending on the version of the model used. Any text exceeding that length is essentially in a completely separate universe with no relationship to the preceding text. The problem with this is that the model cannot then accurately answers inquiries that rely on the relationship between text that appears before this context cutoff and text that appears after it. The results CoCounsel presents often repeat themselves in a way that makes this blindness obvious. CoCounsel, like other AI document analysis tools, likely tries to get around this inherent limitation by using "embeddings," which are basically summaries of shorter text passages, e.g., paragraphs, that are created and accessed on the fly to save space within that context window. This has a different problem: Lost details. A summary is inherently lossy in that it cannot contain all of the information in the original work. When every fact matters, missing something due to an invisible technical background process necessary to "fit" the text into the AI model is likely unacceptable.
- I am a practicing trial attorney, and head of my firm. Both the article and many of the comments below are starry-eyed imaginings that simply misunderstand the legal system and the needs of society. I can't fault the attempt, though; large-language AI is currently being discussed by well-meaning lawyers. My practice works the fair and honest way: overwhelmingly by contingency fee- and all similar attorneys already have a fierce incentive to minimize costs, hours and expense to clients. What this kind of AI will do is vastly improve online legal research sites (like Lexis-Nexis or Westlaw) enabling legal professionals who may not be able to put a logic-based search together to have a machine do it. I will certainly help with massive cases like some class actions. But the ability to search massive document stores to find needles in haystacks is already accomplished with current computers and software.
How will Language Modelers like ChatGPT Affect Occupations and Industries? Ed Felten (Princeton) Manav Raj (University of Pennsylvania) Robert Seamans (New York University)
https://arxiv.org/pdf/2303.01157.pdf
- AI Occupational Exposure (AIOE) measure and used this measure to identify which occupations, industries and geographies are most exposed to AI.
- We find that the top occupations affected include telemarketers and a variety of post-secondary teachers such as English language and literature, foreign language and literature, and history teachers. We also find the top industries exposed to advances in language modeling are legal services and securities, commodities, and investments. We also find a positive and statistically significant correlation between an occupation’s mean or median wage and our measure of exposure to AI language modeling.
Goldman Sachs Article: https://www.ansa.it/documents/1680080409454_ert.pdf
- generative AI technologies currently in focus, such as ChatGPT, DALL-E, and LaMDA, are distinguished by three main characteristics: 1) their generalized rather than specialized use cases, 2) their ability to generate novel, human-like output rather than merely describe or interpret existing information, and 3) their approachable interfaces that both understand and respond with natural language, images, audio, and video.
- Generative AI’s ability to 1) generate new content that is indistinguishable from human-created output and 2) break down communication barriers between humans and machines reflects a major advancement with potentially large macroeconomic effects.
HBR: https://hbr.org/2022/12/chatgpt-and-how-ai-disrupts-industries
https://www.wired.co.uk/article/generative-ai-is-coming-for-the-lawyers
- According to Harvey, one in four at Allen & Overy’s team of lawyers now uses the AI platform every day, with 80 percent using it once a month or more. Other large law firms are starting to adopt the platform too, the company says.
- Generative AI is having a cultural and commercial moment, being touted as the future of search , sparking legal disputes over copyright , and causing panic in schools and universities .
- “Legal applications such as contract, conveyancing, or license generation are actually a relatively safe area in which to employ ChatGPT and its cousins,” says Lilian Edwards, professor of law, innovation, and society at Newcastle University. “Automated legal document generation has been a growth area for decades, even in rule-based tech days, because law firms can draw on large amounts of highly standardized templates and precedent banks to scaffold document generation, making the results far more predictable than with most free text outputs.”
- AI is likely to remain used for entry-level work, says Daniel Sereduick, a data protection lawyer based in Paris, France. “Legal document drafting can be a very labor-intensive task that AI seems to be able to grasp quite well. Contracts, policies, and other legal documents tend to be normative, so AI's capabilities in gathering and synthesizing information can do a lot of heavy lifting.”
- Sereduick says that while the outputs from legal AI will need careful monitoring, the inputs could be equally challenging to manage. “Data submitted into an AI may become part of the data model and/or training data, and this would very likely violate the confidentiality obligations to clients and individuals’ data protection and privacy rights,” he says.