Artificial Intelligence and the Legal Field

There have been a number of technological developments that have impacted the legal field in recent years. The sheer speed at which the internet has expanded has had a massive impact on how humanity engages with the world and with one another. Almost every aspect of the way that we live our lives has changed in the past 50+ years, and the legal profession is no exception to that change. While it may move slower than other professions, the legal field has nevertheless been reshaped by the development of tools like email and online databases, which have fundamentally changed how communication and research function. There are still many situations where brick-and-mortar research is required, and where face-to-face communication is vital, however much of what used to be done physically is now taking place online. Nevertheless, there has always been a need for the human element alongside these technological advances. In recent years, however, many people have begun to believe that is soon to change as the prevalence and use of Artificial Intelligence (AI) increases. 

Artificial Intelligence has been in the human consciousness for millennia. As far back as the 8^th Century BC the ancient Greeks told tales of the Αυτοματονες, or Automatons, animate statues crafted by the god Hephaistos and the Athenian craftsman Daidalos, many of whom were able to think and act as humans did, despite being created from metal.  This has continued into modern-day storytelling in popular media such as the droids in Star Wars or Tony Stark’s AIs such as J.A.R.V.I.S. and F.R.I.D.A.Y. in the Marvel comics and visual media. These advanced AIs remain in the realm of science fiction and fantasy; however, humanity has in recent years managed to develop AIs that are also incredibly powerful.

As of March 2025, all AIs in existence are what can be referred to as Artificial Narrow Intelligence, also known as “Weak AI” or “Narrow AI.” All other forms of AI such as Artificial General Intelligence (AGI or “Strong AI”) and artificial superintelligence (Super AI) are purely theoretical. Narrow AI can be trained to perform tasks, often faster and better than a human mind can. That said, there are limits to Narrow AI, as it cannot perform tasks outside of its parameters. Generative AI (GenAI), while impressively powerful, is still a subset of Narrow AI, as its creations depend on having massive data sets to learn from, and that learning is restricted to those data sets. If GenAI could use those data sets and what it has learned from them to accomplish new tasks in different contexts without the need for humans to train the underlying models, that would be an example of AGI. However, AGI, which indicates an AI that can learn and perform any intellectual task that a human can, along with Super AI, which refers to an AI that can learn and perform any intellectual task better than a human can, remains purely theoretical.

Even though we only have access to  Narrow AI (including the increased prevalence of GenAI), those innovations have had a strong impact on the legal profession and will continue to do so. One innovation that is particularly relevant to the legal profession is the development of natural language processing (NLP). As a linguistics major in college, I became familiar with NLP, which relies on computational linguistics, statistical modeling, machine learning, and deep learning to enable computers and other digital devices to recognize, understand, and generate text and speech. NLP is so present in our daily lives that many people hardly question their presence anymore—among its many uses, it is a large part of how search engines such as Google or Bing answer common queries and is the backbone of programs such as Google translate, Siri, and Alexa. NLP can be used in the context of the legal field to fully or partially automate tasks such as document analysis and review, which can drastically reduce the time that lawyers spend on manual research and document review. 

One place where NLP and GenAI as a whole have changed the game is when it comes to E-Discovery, as AI can be used to quickly sort through massive amounts of electronic data to identify relevant evidence for litigation. As we learned in the legal technologies module, Electronic Discovery is becoming increasingly important as a form of research in legal cases. Nevertheless, when making requests for electronic discovery, the burden test of Fed. R. Civ. P. 26 makes things difficult for the plaintiff, as a defendant may object that there is a significant cost on their part when it comes to electronic discovery. However, the increase in capabilities of AI and tools such as machine learning have drastically reduced the need for human input for some tasks related to discovery. For example,  synthesizing data that could be requested by a plaintiff, and so, depending on the scale of the request, the defendant’s grounds for objection could be drastically reduced. 

AI is present in so many small ways that people rarely think about anymore, but which have a strong impact on everyday working conditions in the legal field. Autocorrect, for example, is maintained by AI and NLP, and yet is something that is completely normalized in society. Many email providers will have a warning pop up when someone tries to send an email that contains the word “attachment” and there is nothing attached to said email—that is AI also. Every Google search relies on AI. Automated tasks such as sending out a recurring bill, calendar reminders for upcoming appointments, auto-saving documents, searching a document for keywords, and generating standard documents from a template, are all things that are done every day by millions of people, including active legal professionals, and yet often aren’t thought of as “AI.”[1]

AI is often branded as something that will streamline processes, reduce costs, and improve accessibility. Nevertheless, there are many dangers present in AI that do not get nearly enough attention. Narrow AI—the only AI we have access to at present—can never fully replace a human, although many carry the false belief that it could. GenAI is completely dependent on its training models, which are often sourced via illegal data mining that does not credit or compensate those who created the original content.  Furthermore, because everything that GenAI creates is dependent on the work of others, if there are not enough humans injecting new information into the data sets everything will become derivative, and innovation will stagnate. Moreover, there are additional dangers when it comes to how the privacy of clients is upheld when their data is being processed by AI—many GenAI models will incorporate the data they are analyzing into their own data sets, and so strict guidelines need to be upheld in that regard. There are nuances to data analysis and connections that can only be made by a human brain and with human creativity, and presently AI simply cannot compensate for that lack. Unfortunately, some employers have been prioritizing AI over the learning and cultivation of human talent and skill. This has resulted in many people losing their employment.

In sum, AI is inevitably here to stay and has had and will continue to have a lasting impact on all aspects of life, and the legal field is no exception. The exact direction it will take us in is dependent not only on the developers who create these massive AI models but on us as legal professionals as we decide to what extent we will allow its use in our field. 

[1] To be clear, not all of these processes are always governed by AI—however where many processes may have originally relied on other types of machine intelligence, the prevalence of AI has shifted the backend of how they  are conducted.

Bibliography

Atsma, A.J. (no date) Automotones, Theoi Project. Available at: https://www.theoi.com/Ther/Automotones.html (Accessed: 03 March 2025).

Holdsworth, J. (2025) What is NLP (Natural Language Processing)?, IBM. Edited by C. Stryker. Available at: https://www.ibm.com/think/topics/natural-language-processing (Accessed: 03 March 2025).

IBM Data and AI Team (2025) Types of artificial intelligence, IBM. Available at: https://www.ibm.com/think/topics/artificial-intelligence-types (Accessed: 03 March 2025).

Note: The original version of this essay was written in March 2025 as the final assignment for my Legal Technologies module while completing my Paralegal Studies Certificate at Boston University.