Generative AI in Law: Generating a New Course for Law?
- Daivik Soti
- Aug 23, 2024
- 9 min read
Updated: Aug 24, 2024
This article explores the transformation in legal practice due to Artificial Intelligence (AI), focusing on its integration into law schools, law firms, and courts. It reviews the positive impacts, such as increased efficiency and cost savings, alongside challenges like plagiarism, ethical dilemmas, and inaccuracies in AI-generated legal advice. Highlighting both advancements and risks, the article emphasizes the need for careful use and regulation of AI in the legal field to uphold professional standards and ethics.
*Daivik Soti
INTRODUCTION
Twenty years ago, from today, one may find a lawyer and/or his apprentice in his chamber or in a library, sifting through books, journals and reportable like All India Reportable to try and find relevant case laws for their motions. Now, such a sight is rare; nigh impossible to find in the modern legal fraternity.
The digitisation of law was a much-needed breath of fresh air for one of the oldest professions in the modern world, and now, the legal fraternity is anticipating, or bracing, depending on who you ask for the second age of digitisation of law, namely, integration of law with Artificial Intelligence (AI), with generative AI leading the charge front and centre.
This article aims to look into the good, the bad, and the hallucinatory of artificial intelligence in law as it stands today. I would like to restrict myself to the limited scope of how Artificial Intelligence impacts law in this day and age, and not predict or forecast the impact it may or may not have on law. I would like to conclude with a cautionary note on how a lawyer should use this technology, touching upon the field of legal ethics.
In late 2022, OpenAI launched their highly anticipated chatbot ChatGPT, the first commercially available chatbot capable of answering text prompts with long-form answers. Harvard Business Review called this “the tipping point of AI”[1], while Reuters published a report on how ChatGPT achieved the milestone of a 100 million monthly active users within 2 months of his launch, making it the fastest growing consumer application by a mile.[2]
1. ChatGPT’s Transition in Legal Academia from Scepticism to Integration
The first mass use of ChatGPT in law can be seen in the nascent stage of law, i.e., law schools. Law students were the first to jump on the bandwagon of generative AI, using ChatGPT in order to help with their cumbersome projects, or summarise some of the longest judgements pronounced by various courts of law in small, easy to understand paragraphs.
The response of Law Schools around the world vis-à-vis generative AI, to oversimplify, has been mixed. Initially, law schools were apprehensive and even critical of AI in legal education, with well found concerns about plagiarism and IP violation. Professors were quick to warn their pupils against AI use in assignment, quickly followed by formal and restrictive AI policies by schools like the Chicago Law School, whose Law School Policy on Generative AI[3] was quick to outright ban use of AI in examinations, attribute AI text as AI’s and not their own, and empowering the instructor with the final say on how AI can and cannot be used in their subject.
Eventually, law schools realised that AI is the next big thing in technology and by addendum, technology law. Schools like the Harvard Law School (Initiative on Artificial Intelligence and the Law)[4] and Vanderbilt Law School (AI Legal Lab) were quick to launch their courses on Artificial Intelligence Law, with such courses furthering legal professionals’ understanding of AI and analyse how these generative tools affect the access to justice and practice of law.
Going a step further, we analyse the impact of AI in the field of law, both at the courtrooms and at corporate law firms. The integration of generative AI was slow in these fields at first, but seeing the seemingly unlimited potential of AI, the legal fraternity was quick to jump on the bandwagon as well which has perhaps forever changed the course and trajectory of our profession and may even change our understanding and definition of law and legal practice.
2. The Impact of AI on Law Firms and Courts
Law firms invested heavily in developing unique law focused versions of ChatGPT chatbots and other AIs, and prima facie, their investments have borne great fruit. AI is now used for predictive litigation analysis, electronic predictive coding to categorise and analyse volumes of text, calculating damages in a suit for breach of contract, managing contracts to help pinpoint performance clauses and scope of rights and liabilities, and even going as far as to analyse the electronic communications of a party to analyse wrong doings. Such changes have saved thousands of hours for lawyers and paralegals, in turn saving millions of dollars for large law firms.
It must be noted however, that AI has not only just benefitted big ‘suits’ in the profession. Litigating lawyers, at all stages of litigation have many tools at their disposal to make their workload a bit easier and efficient. Legal research can be expedited with the help of summarising AI and drafts of motions for contracts, regulatory filings, wills, trusts, patent specifications, affidavits, articles of incorporation, and other documents can be produced using generative AI, which rarely creates any of these papers from start; instead, it uses a mix of input data, template databases, and prompts to generate them.
Even the judiciary, an organ of the government notoriously known for its resistance to change, and outside intervention has surprisingly (or unsurprisingly, depending on who you ask) has embraced Artificial Intelligence to make work in the courtrooms a bit easier and a lot more efficient. The Brazilian Supreme Court’s adoption of the VICTOR system is perhaps the most famous example of an Apex Court’s active use of AI in its day-to-day activities. VICTOR is being used to deal with thousands of appeals by facilitating the quick identification of those meeting the “general repercussion” criterion.
Automated initial sorting and analysis of appeals by VICTOR help the court better manage its caseload, and it ensures that only such cases as have important legal implications will be reviewed. The Indian Apex Court, while not adopting AI to the standard of its Latin and European peers, has experimented with an indigenous sophisticated software known as SUVAS for the translation of a large bank of legal documents from English into ten vernacular languages. One can only wait and watch to see how the Indian and the global justice community adopts AI in its practice.
3. Conclusion: A Cautionary Path
It can be said without a shadow of the doubt that the overall impact of AI on the legal field, for all means and purposes, is overall positive. Law students and lawyers alike have benefitted greatly from generative and summarising AI, allowing greater human focus on analytic rather than clerical tasks. This allows for lesser workload on overburdened and underpaid interns and employees, and better legal education in law schools. However, the use of AI, and its future implications have raised some well-founded questions and concerns.
3.1 Plagiarism
Firstly, the nuisance of plagiarism has arose to new heights in this era of generative AI. Not just law students, but even law academicians have used AI in order to cut corners and publish more academia. This has resulted in an overall decline in the quality of both legal education and legal publication and seen responses like Turnitin’s AI detection and improved plagiarism checker. There have also been calls for an outright ban on use of AI in law school assignments, and written undertakings from legal academicians on non-use of AI in their publishing.
3.2 Attorney-Client Privilege
Moreover, generative nature of AI’s has some inherent ethical dilemmas lawyers must navigate through. I would like to discuss the impact on Attorney-Client Privilege by use of generative AI. There are two inherent problems here. Firstly, the online database these generative AI use for legal purposes contain not only judgements, but also case filings and other court documents. The concept of privileged communication is challenged at its very roots here. The original client of such cases would have of course contained sensitive information in the filings, and such information being made public by such databases impacts both the privacy of the parties to the suit and the confidentiality of privileged communication.[5]
3.3 Legal Advice
These generative models are still at a nascent stage, with factual inaccuracies and incorrect interpretation of inputted information being a very real risk. Hence, the standards of checks and vigilance entailed by lawyers and law firms should be much higher, so as to avoid such technological mistakes having an adverse effect on the case of their client and the concept of justice. From a legal-ethical perspective, such errors by generative AI would also be classified as improper legal advice; since the Court is much more likely to view this as a case of negligence on part of the lawyer.[6]
3.4 Hallucinations in AI
The term "generative AI hallucinations" describes inaccurate or deceptive outcomes produced by AI models. Numerous things, like as biases in the data used to train the model or insufficient training data that lead to erroneous assumptions made by the model, might result in these errors. The problems and risks of hallucinations in legal AI cannot be described as anything but grave.[7]
The main basis of incorrect legal advice is the inherent hallucinatory nature of generative AI models. Law is a complicated, diverse and sensitive to detail subject matter. Generative AI models are trained to cater to a broad range of general topics. These nascent AI models, when used to deal with extremely specific fields of law and individual cases, are prone to overgeneralising and generation of their outputs by using their entire database, which leads to ‘data spills and hallucinations in the output.
3.5 Legal Chatbots
Lastly, I would like to address the problem of legal chatbots, being developed by law firms and law startups based on the generative AI models of companies like Law Droid and Chat-fuel by Facebook. Both these companies deal with creating personalised chatbots for websites and companies, which can be used by both their employees and their clients.
Law startups, both in India and abroad have used these services to create end user chatbots which clear legal doubts and give a degree of legal advice to the common populace. The problem with websites like these are multilateral. Firstly, the more niche an AI is, the more limited its database is on which it is trained on and hence suffers from both limited knowledge and increased bias. Hence, users are given wrong legal advice or advice which is filled with bias. Moreover, the advice given by these chatbots is unmoderated and unfiltered, with no system of checks and proofreads, which makes such advice all the more dangerous.
As lawyers, it is a dire question of professional ethics that we ensure that any advice we give, may it be directly or through intermediaries like these AI chatbots, must be well researched and free from bias and negligence, otherwise, AI would strike at the core of legal ethics.
A Brainstormer!
We hope you found this article informative and insightful. We would like to invite you to reflect on some of the questions that arise from the use of generative AI in the legal field, such as:
How can we ensure the accuracy and reliability of AI-generated legal advice and documents?
-How can we protect the confidentiality and privacy of clients and parties involved in legal cases that are used as data sources for AI models?
How can we prevent the misuse and abuse of AI for plagiarism and fraud in legal education and publication?
How can we balance the benefits of AI for enhancing legal efficiency and productivity with the ethical and professional responsibilities of lawyers?
If you have any feedback or opinions on these or any other related issues, we would love to hear from you. You can write to us at team@legalverse.in and we may publish your views on our platform. Thank you for reading.
*Daivik Soti is a 5th-year Constitutional Law (Hons.) student at National Law University, Jodhpur.
________________________________________________________________________________
[1] Ethan Mollick, ChatGPT Is a Tipping Point for AI, HARV. BUS. REV. (Dec. 14, 2022), The same can be accessed from here: https://hbr.org/2022/12/chatgpt-is-a-tipping-point-for-ai.
[2] Krystal Hu, ChatGPT Sets Record for Fastest-Growing User Base - Analyst Note, REUTERS (Feb. 2, 2023), The same can be accessed from here: https://www.reuters.com/technology/chatgpt-sets-record-fastest-growing-userbase-analyst-note-2023-02-01/
[3] Law School Policy on Generative AI, University of Chicago, The same can be accessed from here: https://www.law.uchicago.edu/students/handbook/academicmatters/generative-ai
[4] Harvard Launches New Initiative to Better Understand—and Shape—the Future of AI, Cassander Coyer (July 18, 2023), The same can be accessed from here: https://www.law.com/legaltechnews/2023/07/18/harvard-launches-new-initiative-to-better-understand-and-shape-the-future-of-ai/?slreturn=20240723060131
[5] Serena Villata et al., Thirty Years of Artificial Intelligence and Law: The Third Decade, 30 A.I. & L. 561, 561 (2022), The same can be accessed from here: https://link.springer.com/article/10.1007/s10506-022-09327-6
[6] John Armour & Mari Sako, AI-Enabled Business Models in Legal Services: From Traditional Law Firms to Next-Generation Law Companies? 7 J. PROS. & ORG. 27 (2020), The same can be accessed from here: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3418810
[7] Jonathan H. Choi & Daniel Schwarcz, AI Assistance in Legal Analysis: An Empirical Study, 73 J. LEGAL EDUC, The same can be accessed from here: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4539836
Comments