A few decades ago, legal practitioners spent countless hours amidst towering piles of physical text, carefully drafting handwritten briefs for their clients. Since the late twentieth century, this old-fashioned practice has become modernized, beginning with the transition from typewriters to computers, the introduction of online legal research databases, and recently with the use of online court hearings during the COVID-19 pandemic Today, the legal profession is in the midst of its most recent technological booster, as artificial intelligence (“AI”) has become mainstream and begins its inevitable assimilation with the industry. However, AI technology presents different consequences and concerns due to its rapid development and a disconcerting fear of what it could ultimately mean for lawyers around the world. Most recently, news of the first robot scheduled to appear in court received so much backlash, that its creator postponed the appearance due to fear of criminal prosecution. Moreover, one of ChatGPT’s latest iterations claims the ability to pass the Uniform Bar Exam (“UBE”), achieving a score in the 90th percentile of human test-takers.
As the use of AI in legal settings raises significant ethical, legal, and technical concerns, it is imperative to provide a comprehensive account of the potential issues in order to effectively address them. Certainly, there is a need for government intervention, to provide a democratically selected framework for AI. However, as with all other forms of technology, AI is a tool, so lawyers who adopt and perfect its early usage will surely find themselves ahead of the curve on what is likely to become the new norm for the legal industry.
Section 1 – Benefit to Clients – Access to Legal Services
In a nation with more than 330 million people, there are only 1.3 million lawyers,The Unmet Need for Legal Aid, LSC Am.’s Partner for Equal Just., https://www.lsc.gov/about-lsc/what-legal-aid/unmet-need-legal-aid [https://perma.cc/TY7N-E47T].[/efn_note] Technology, and the internet, has played a role in reversing that discrepancy. One example is the rise of companies like Nolo and LegalZoom, which are designed to help users generate legal documents without the need, or the cost, of an actual lawyer. Some argue that this access is counterproductive and, in fact, not in the best interests of individual clients. After all, there is no personal connection in this situation, and certainly no attorney-client dynamic, much less is there any legal screening process.
Quite plainly, the market these companies occupy is one that rightly belongs to and is safeguarded by those who attended law school and passed the relevant bar exam. However, these companies are able to exist because of the low costs they offer. Instead, imagine a future where legal professionals harness the power of AI to draft legal documents adroitly and affordably, with clients able to conveniently provide necessary information via secure online portals and opt for cost-effective review services before receiving the final product. Such a business model would only cost clients slightly more than the services provided by Nolo and LegalZoom. For clients seeking the invaluable combination of legal acumen and empathetic guidance in their representation, the slightly higher cost of retaining a human lawyer is a justifiable investment, as it affords a uniquely personalized and nuanced approach that transcends the limitations of automated chatbots. As clients engage in question-and-answer sessions, the potential for increased billable hours becomes more likely, exemplifying the “foot-in-the-door” sales tactic.
This hypothetical is likely achievable with the assimilation of AI, as its use will allow lawyers to do work more efficiently, thereby spending less hours per client. Less time spent with each individual client will likely allow firms more time to acquire new clientele, leading to greater overall profits and larger referral pools. Furthermore, as the cost of legal services adjust, it is likely that more clients will look to acquire legal assistance than before because of this increased accessibility, affordability, and retained human interaction.
Issue 2 – Benefits to Lawyers – With Great Power Comes Great Responsibility
Today, almost all legal research is done using online legal databases like WestLaw and Lexis. However, these services are amplified with the integration of AI, by companies like ROSS and now ChatGPT. Able to sift through thousands of legal decisions in minutes, AI can determine what legal precedents have been cited most often to produce a desired result, it can help determine what cases an opponent may cite to, and, perhaps as a result, help guide precedent.
A commonly used form of AI already exists in Casetext, CoCounsel’s AI legal assistant. Casetext purports to be used by over 10,000 law firms, including firms of all sizes, from solo practitioners to BigLaw firms. Among the reviews on their website are lawyers swearing by the technological advancement, their comments praising the ability of CoCounsel to do work that would require several attorneys working countless hours (e.g. reviewing boxes of documents), to the ability to spend this time on better things like developing greater attorney-client relations. This further bolsters the benefits of the hypothetical discussed in Section 1, through the use of AI, lawyers are able to focus on more human-demanding tasks like creating nuanced legal arguments, giving bigger clients more attention, and building their brand.
Realizing that the use of AI in law needs some guidelines, the ABA initially stated the following:
[T]he American Bar Association urges courts and lawyers to address the emerging ethical and legal issues related to the usage of artificial intelligence (“AI”) in the practice of law including (1) bias, explainability, and transparency of automated decisions made by AI; (2) ethical and beneficial usage of AI; and (3) controls and oversight of AI and the vendors that provide AI.
Essentially, the ABA adopts the usage of AI but makes it clear that “[l]awyers must not only advise clients [of its usage] but also contend with legal, ethical, business, and malpractice risks involved in deploying AI in their practices.” Furthermore, the ABA recently adopted a new resolution holding individuals and organizations that develop, deploy, and use AI systems and capabilities accountable for the consequences caused by the use of AI systems, unless they have taken reasonable steps to mitigate against that harm or injury.
This framework has been endorsed by many who currently incorporate AI, including the firm Allen & Overy, with guidance from David Wakeling, who witnessed the use of AI blossom to the point that “one in four [of their attorneys] . . . now uses the AI platform every day.” The policy adopted by that firm is quite clear, as Mr. Wakeling articulated, “[lawyers] must validate everything coming out of the system. [They] have to check everything,” to avoid any potential issues with its usage. Two points become immediately apparent; first, lawyers are necessary to serve as the final barrier that AI-produced work must overcome before it can be sent to clients; and, second, as with all forms of technology, AI will pose its own unique risks.
Section 3 – Risks, Competency and the Unauthorized Practice of Law
One potential risk associated with AI are hallucinations, which in this context, refers to mistakes in AI-generated text that are “semantically or syntactically plausible but are in fact incorrect or nonsensical.” This concern stems from the underlying premise of AI, in that it is scouring the internet to provide feedback based on a submitted prompt. Two areas of concern are presented in this scenario, first, the submitted prompt, or issue, must be specific and on point, and, second, as we have seen the internet is rife with misinformation that will ultimately influence AI’s ability. After all, fundamentally, AI does not know what it does not know, thus, as the aptly named hallucination implies, it is visualizing something that is not actually there.
However, the solution to this problem seems to be trial by users, supervision by lawyers, and a hopeful sentiment that AI will not be corrupted by misinformation like some believe has already happened to social media. Since lawyers who decide to use AI will act as final arbiters, these concerns demand lawyer competency as required by the Model Rules of Professional Conduct, which incorporates AI in its broad “relevant technology” language. Summarily, lawyers must maintain the “requisite knowledge and skill,” while remaining current on the benefits and risks associated with AI. This new liability may be proactively dealt with by creating and implementing strict guidelines for its usage by firm employees, including a revisionary system that certifies substantive quality.
The use of AI also raises significant privacy concerns, as it necessitates uploading what could be privileged client information to third parties platforms. Information may be stored for extended periods of time, increasing the risk of unauthorized access and potential breaches. This concern directly implicates the attorney-client privilege, a foundational aspect of legal representation.Several countries have begun to address the protection of personal data, from the European Union’s AI Act, to Italy’s Data Protection Agency, and the United States Office of Science and Technology Policy (“WHOSTP”). In fact, Italy recently instituted a temporary ban on ChatGPT, demanding additional information regarding personal data collection.Importantly, this type of governmental guidance has been requested by AI companies, specifically from OpenAI’s (developer of ChatGPT) CEO. Until further oversight is implemented, lawyers are in the perfect position to address this concern with clients, and in doing so, satisfy ABA requirements. To this end, law firms that currently implement AI, do so by not using a client’s personal information. Due diligence further requires firms that utilize AI to implement strong data privacy safeguards, including scrutinizing AI platforms before utilizing their service, informing firm personnel of strict guidelines, data encryption, and giving clients the option to decline the use of AI services. In fact, the latter may be used as a way to gain clients by advocating the non-use of AI platforms for clients who may prefer the lack of any third-party knowledge.
Lastly, to what extent is the use of AI an unauthorized practice of law? Model Rule 5.5(b) explains what the ABA defines as the unauthorized practice of law, but courts have yet to really establish legal precedent in the area of AI’s usage. However, several recent decisions are informative. In the case of Lola v. Skadden, Arps, Slate, Meagher & Flom LLP, the court held that document review was not per se within North Carolina’s definition of practicing law. Specifically, the court explained that practicing law includes exercising “at least a modicum of independent legal judgment.” Furthermore, in Janson v. LegalZoom.com, Inc., the court held that filling out blank forms – like those provided on websites like LegalZoom – is not, “in and of itself the unauthorized practice of law.” However, it is worth noting that this case was ultimately settled after the court denied the defendant’s summary judgment motion because LegalZoom did more than simply provide templates, it essentially took control, turning its “self-help kit” into more of a “we’ll do it for you” system.
Furthermore, the upcoming ruling in Gonzalez v. Google LLC has the potential to impact the use of AI in legal settings, specifically in relation to Section 230 of the Communications Decency Act.Depending on the Supreme Court’s decision, online platforms may become more exposed to legal liability for user-generated content. Alternatively, if the broad interpretation of Section 230 immunity is upheld, it would provide a legal foundation for the continued advancement and implementation of AI-powered systems in legal settings. The core argument of this case is that certain websites (specifically, Google), through algorithms actively partake in the dissemination of information when they recommend content to their users. Surely, if search engines do this now, AI like ChatGPT is even more exposed to liability, thus this decision will play a key role in that future determination.
Conclusion
AI’s proliferation may be scary for lawyers and certainly for law students (as some may fear a shrinking job pool), however there are certain qualities that AI cannot attain – at least not yet. Those qualities are what make humans, human; and, what makes lawyers necessary. Therefore, the lawyers who perform well in creativity, judgment, empathy, and adaptability will remain highly sought after. In fact, some firms have recently focused on this notion, and have begun to change the manner in which they select employees. It is as though they have realized that this integration is the future of the profession, or perhaps they have perceived a competitive advantage. After all, it is easy to imagine a situation where lawyers and clients who utilize AI will be in a better position than lawyers and clients who do not.
Ultimately, the role of AI, much like any other technological advancement, has its own advantages and disadvantages. It is an exciting opportunity for skilled attorneys to utilize technology to enhance their work and better serve clients, thus, if employed with the proper expertise, it can undoubtedly serve as a valuable tool. The legal industry can improve access to justice, increase efficiency, generate greater profits, and better serve clients. However, if implemented carelessly, it may end up doing more harm than good. The key is to recognize the limitations of AI and to use it as a complementary tool to lawyer expertise, rather than a replacement. The tide has already begun shifting, and due to its wide reach, lawyers can now receive CLE credit from attending sessions discussing the implications of AI in the law, with some states even mandating this training.
While this episode of the Twilight Zone is only just beginning, so long as the plaintiffs, defendants, judges and juries are humans, there will always be a need for human lawyers. The assimilation of AI and the legal field will continue the latter’s growth, with firms that take advantage of the benefits of AI leading the charge as they are able to offer affordable and better legal services, which thus far, has led to growth and success of the legal field.
Ksenia Khlystova-Gowda is a Second Year Law Student at the Benjamin N. Cardozo School of Law and a Staff Editor at the Cardozo Arts & Entertainment Law Journal. Ksenia is interested in family, privacy, and intellectual property law.