Written by: Micah Drayton
AI has impacted industries from the medical to education, and the legal world is no different. Explore how AI is advancing the legal realm.
AI has impacted industries from the medical to education, and the legal world is no different. But as I take a closer look at the impact of AI in my part of the legal world, I see it as useful but limited. Our firm uses some AI products to generate predictions regarding the examination of patent applications, such as what classification of invention or technology type the application will receive at the USPTO. This is a useful way to check that we are communicating important aspects of inventions effectively and helps us predict how the examination is likely to proceed. It is not perfectly accurate, but is accurate enough to be helpful and requires a lot less effort than the more manual approach we used before.
We have also explored the ability of AI to produce initial drafts of legal documents. While it is true that working off an existing draft can often be more efficient than writing it from scratch, a pretty crucial step in writing a legal document is to think about the structure of the document and use that thinking to organize the writing. Language-based AI is especially challenging to simulate, and the industry might not be there yet. Currently, AI does not help very much with certain kinds of legal writing.
AI interests me both as the subject of the inventions I work with and as a potential consumer of AI. In the former case, it is clearly becoming increasingly ubiquitous in everyday life. Much like the proliferation of microcontrollers and other low-power processor devices in the Internet of Things, AI has increasingly been working in the background and under the hood in all kinds of mundane, everyday activities, frequently without the awareness of the people who benefit from its use.
In the legal profession, I am excited about the ability of AI to enhance legal research by locating documents or passages of relevance to a topic of interest. This is something that conventional keyword searching does not perform adequately, as the semantic relationships to be discovered are more subtle. Use of vectors to find related terminology is currently enhancing some of the more sophisticated search algorithms, but what will really help will be algorithms that can search by concepts as opposed to semantics. These would enable you to, e.g., find a case that (a) depends on the controlling precedent at issue and (b) resembles an instant case according to more abstract or conceptual levels of similarity, rather than merely matching up words and synonyms.
As with every other advance in automation, it is not clear yet what impact this will have on employment and careers. I’d say a big concern is that AI could be used to produce superficially passable legal work at discount prices. Such output might appear, particularly to customers who are not well-versed in the law, to perform its function adequately, but may have hidden flaws that, absent responsible review by skilled practitioners, might not be discovered until it is too late. What, for instance, will stop an unscrupulous lawyer from contracting to produce many wills or trust documents, use an AI platform to crank them out with little or no oversight, and pass the result off as his or her work? In that case, many of the documents’ efficacy would not become apparent for years or decades and could lead to contention or loss long after our enterprising huckster has retired to a remote island.
Much like the AI that runs automated phone systems, AI-produced legal instruments may go through a period in which they adequately address the most common or typical issues, but fail to address, or perhaps miss entirely, the situations that require closer attention. This is not the sort of thing people should gamble with, and it would be hard to prevent before it was able to do considerable damage. While AI is an exciting and possible time-saving technology there are still limitations on how it can be used, especially in the legal community. As AI improves, we will see people in favor of more adoption of AI, but only if solutions to some related issues improve as well.
This publication is distributed with the understanding that the author, publisher, and distributor of this publication and/or any linked publication are not rendering legal, accounting, or other professional advice or opinions on specific facts or matters and, accordingly, assume no liability whatsoever in connection with its use. Pursuant to applicable rules of professional conduct, portions of this publication may constitute Attorney Advertising.