Introduction
The rapid proliferation of sophisticated generative artificial intelligence (GenAI), exemplified by models like OpenAI’s ChatGPT, has presented a profound and disruptive challenge to higher education. Within the discipline of law, this disruption is particularly acute. Legal education is predicated on the development of core skills such as critical analysis, precise legal reasoning, and the assimilation of complex, authoritative sources—all of which appear to be capabilities that GenAI can simulate. This has forced legal academics and institutions into a difficult position, prompting a critical debate encapsulated by the question: should GenAI be treated as a legitimate study aid or as a fundamental risk to academic integrity? To frame this as a simple binary choice, however, is to mischaracterise the nature of the technology and its potential impact. An outright ban is both Luddite and largely unenforceable, while unregulated adoption risks a catastrophic dilution of academic standards and professional competence.
This essay will argue that the ‘aid versus risk’ framing is a false dichotomy. GenAI is simultaneously a significant academic integrity risk and a potentially transformative study aid. Therefore, the appropriate response from UK law schools is not one of prohibition, but one of proactive and critical engagement. It will be contended that legal education must develop a nuanced framework that mitigates the risks while harnessing the benefits. This framework necessitates three core components: first, the establishment of clear and transparent institutional policies on permissible use; second, a deliberate evolution in pedagogical practices and assessment design; and third, the integration of ‘AI literacy’ as a foundational skill for the modern law student. By treating GenAI as a tool to be mastered rather than a threat to be vanquished, legal education can not only safeguard its integrity but also better prepare its graduates for the realities of a technologically advanced legal profession.
The Spectre of Generative AI: A Fundamental Risk to Academic Integrity
The most immediate and widely discussed threat posed by GenAI is to academic integrity, manifesting in several distinct but interrelated ways. The primary concern is that it represents a new and insidious vector for plagiarism and academic misconduct, fundamentally challenging traditional notions of authorship and originality. Unlike rudimentary copy-and-paste plagiarism from existing sources, GenAI tools produce novel, seemingly original text in response to a user’s prompt (QAA, 2023a). This makes detection by conventional software like Turnitin more challenging, as the generated text does not directly match any single source in the database. The ease and speed with which a student can generate a passable, if mediocre, essay response undermines the entire purpose of assessment as a measure of individual knowledge, understanding, and analytical skill. As Draper and Newton (2023) argue, this capability goes to the heart of what constitutes academic work, creating an “authenticity gap” where the submitted work may bear no relation to the student’s own cognitive effort.
Beyond the issue of authorship, a more pernicious risk in the legal context is the phenomenon of ‘hallucination’. GenAI models are designed to generate plausible-sounding text, but they lack a true understanding of facts or the concept of authoritative sources. This can lead them to invent information, including fabricating case law, statutes, and academic commentary with alarming confidence (Perlman, 2023). The real-world consequences of such uncritical reliance were starkly illustrated in the US case of Mata v Avianca, Inc, where lawyers were sanctioned by the court for submitting a legal brief containing multiple citations to entirely fictitious judicial precedents generated by ChatGPT (No. 22-CV-1461 (PKC), (S.D.N.Y. June 22, 2023)). For a law student, whose entire training is based on the meticulous use of and respect for binding authority, this risk is catastrophic. Submitting an assignment with fabricated cases is not merely poor scholarship; it is a fundamental failure to perform the most basic task of a lawyer and represents a grave breach of academic integrity. It fosters a dangerous disregard for the verification and accuracy that underpins the rule of law itself.
Furthermore, an uncritical over-reliance on GenAI threatens to cause a progressive erosion of the core cognitive skills that legal education is designed to cultivate. The process of reading voluminous and complex judgments, synthesising disparate legal principles, constructing a coherent argument, and articulating it with precision is what builds a lawyer’s mind. If students can outsource the tasks of summarising cases, structuring essays, and even formulating arguments to an AI, they bypass this essential, if arduous, developmental process (Eaton, 2023). This could lead to a generation of law graduates who lack the foundational abilities of critical thinking, analytical reasoning, and complex problem-solving. They might know how to prompt a machine, but they may not have developed the intellectual resilience and rigour required to grapple with novel legal problems where no pre-digested answer exists. This is not just an academic concern; it is a direct threat to the competence of future legal professionals, a matter of central importance to regulatory bodies like the Solicitors Regulation Authority (SRA) and the Bar Standards Board (BSB). Finally, the rise of powerful, subscription-based AI models risks creating a new dimension of educational inequality, where students with the financial means to access premium tools gain an unfair advantage over their peers (Bearman, 2023), further compromising the principles of fairness in assessment.
Harnessing the Machine: Generative AI as a Legitimate Study Aid
Despite the profound risks, to dismiss GenAI solely as a threat is to ignore its significant potential as a legitimate and powerful pedagogical tool. When used thoughtfully and critically, it can augment and enhance the learning experience in ways that traditional resources cannot. One of the most promising applications is its function as a personalised, ‘Socratic’ tutor. A law student struggling with a complex concept, such as the tripartite test for duty of care in negligence or the intricacies of constructive trusts, can ask GenAI to explain it in multiple different ways, request analogies, or generate hypothetical scenarios to test their understanding (Cantatore, 2023). This provides on-demand, individualised learning support that can supplement lectures and seminars, helping to scaffold student understanding and build confidence, particularly for those who may be hesitant to ask questions in a classroom setting.
Secondly, GenAI can serve as a valuable instrument for developing legal writing and drafting skills, provided it is used as a starting point rather than a final producer. A student can use it to help overcome ‘writer’s block’ by generating a basic essay structure, which they must then populate with their own research and analysis. It can act as a sophisticated grammar and style checker, offering suggestions on how to make prose more concise and impactful (Choi, Lee and Lee, 2023). For example, a student could draft a contractual clause and ask an AI to critique it for ambiguity, forcing them to think more carefully about the precision of legal language. In this capacity, the AI is not doing the thinking for the student; rather, it is providing formative feedback that encourages the student to reflect on and improve their own work. This process mirrors how practising lawyers often use templates and boilerplate documents, which they must then adapt and refine to fit the specific needs of their clients.
Thirdly, while its unreliability with case law makes it a hazardous primary research tool, GenAI can be used strategically at the preliminary stages of legal research. A student could use it to gain a broad overview of a legal topic or to identify potential keywords and lines of inquiry to pursue in authoritative legal databases like Westlaw or a reliable case summary site like lawcases.net. The key is to teach students to treat the AI’s output with professional scepticism and to see it as a signpost, not a destination. This process itself is a valuable lesson in the critical evaluation of sources and the absolute necessity of cross-verification—a skill essential for legal practice.
Finally, and perhaps most importantly, engaging with GenAI in law school is essential preparation for the future of the legal profession. As scholars like Richard Susskind (2023) have long argued, technology is fundamentally reshaping the delivery of legal services. Law firms are increasingly adopting AI for tasks such as document review, contract analysis, and legal research. The SRA’s principles require solicitors to maintain their competence, which explicitly includes an understanding of the technology they use (SRA, 2019). Similarly, the BSB’s Professional Statement includes technological proficiency as a competency (BSB, 2016). For law schools to prohibit the use of a technology that is becoming integral to the profession would be a disservice to their students. It would create an artificial schism between legal education and legal practice, leaving graduates ill-equipped for the workplace they are about to enter. Therefore, learning how to use GenAI tools effectively, ethically, and critically is not merely a study aid; it is becoming a necessary professional skill.
Navigating the Dichotomy: A Framework for Responsible Integration
Recognising that GenAI is both risk and opportunity, the central challenge for legal education is one of management and integration, not prohibition. A robust framework is required to navigate this_dichotomy. The first and most fundamental step is the development and dissemination of clear, explicit, and nuanced institutional policies regarding AI use. Outright bans are largely futile due to the difficulty of detection and the ubiquity of the technology. Instead, universities and law schools must move from a position of ambiguity to one of clarity, providing students with precise guidance on what constitutes permissible and impermissible use (QAA, 2023b). For instance, a policy might permit the use of GenAI for brainstorming and proofreading but prohibit the submission of any AI-generated text in an assessed work. It might require students to acknowledge where and how they have used an AI tool, promoting a culture of transparency rather than concealment. Such policies provide the certainty needed for both students and staff to navigate this new terrain.
Secondly, and more transformatively, legal educators must engage in pedagogical adaptation, redesigning assessments to be both more resistant to misconduct and, in some cases, to be inclusive of AI. Assessments that rely on simple information recall or generic essay questions are highly vulnerable to AI. The solution is to shift the focus towards testing higher-order skills that GenAI cannot easily replicate. This could involve an increased use of in-class examinations, oral presentations or vivas, complex problem questions based on novel fact patterns that require deep analogical reasoning, and reflective assignments where students analyse their own learning process (Salmon, 2023).
Concurrently, educators can design “AI-inclusive” assessments. For example, an assignment could ask students to use GenAI to produce a draft legal opinion and then submit a critical evaluation of that draft, identifying its errors, omissions, and weaknesses with reference to primary legal sources. This type of assessment does not test the student’s ability to recall information but rather their capacity for critical judgement, evaluation, and professional scepticism—precisely the value-added skills of a human lawyer in an age of AI. It turns the tool of potential cheating into a tool for teaching critical analysis.
The third pillar of this framework is the explicit teaching of AI literacy as a core competency. Law schools should not assume that students instinctively know how to use these powerful tools responsibly or effectively. Curricula should include instruction on the nature and limitations of large language models, the risks of bias and hallucination, the ethics of using AI in a professional context, and the practical skills of ‘prompt engineering’ to elicit more useful and accurate responses (Kerr and Rilat, 2023). By formally teaching these skills, law schools reframe the issue. The use of AI ceases to be a clandestine activity associated with cheating and becomes a recognised, albeit carefully regulated, part of the modern lawyer’s toolkit. This approach aligns academic training directly with the professional need for technological competence recognised by the SRA and BSB, ensuring that graduates are not only academically sound but also professionally relevant.
Conclusion
The advent of generative AI presents legal education with one of its most significant challenges in decades. Viewing it through the simplistic lens of ‘legitimate study aid versus academic integrity risk’ is a flawed approach that fails to capture the complexity of the issue. The risks to academic integrity—plagiarism, factual inaccuracy, and the erosion of core skills—are undeniable and severe. Yet, the potential of GenAI to act as a powerful pedagogical tool for enhancing understanding, developing skills, and preparing students for a tech-centric legal profession is equally compelling.
This essay has argued that the only tenable path forward is one of critical and strategic integration. A prohibitive stance is not only impractical but educationally regressive. Instead, UK law schools must embrace the challenge by developing a sophisticated, multi-pronged framework. This involves instituting clear policies that replace ambiguity with transparency, fostering pedagogical innovation that prioritises the assessment of higher-order critical skills, and explicitly teaching AI literacy as a fundamental component of a modern legal education. By adopting this approach, institutions can effectively mitigate the academic integrity risks while simultaneously leveraging GenAI to enrich student learning. Ultimately, the goal is not to create lawyers who are insulated from technology, but to cultivate legally-minded, ethically-grounded professionals who know how to master powerful tools, question their outputs, and exercise the uniquely human judgment that will remain the hallmark of a good lawyer, now and in the future.
References
Bearman, M. (2023) ‘A Generative AI Policy for an Equitable, Enabling and Critical Approach’, Postdigital Science and Education, 5(2), pp. 278-282.
BSB (2016) Professional Statement for Barristers. Bar Standards Board.
Cantatore, F. (2023) ‘The Use of Generative Artificial Intelligence (AI) in Legal Research and Writing: A New Challenge for Legal Education’, The Law Teacher, 57(3), pp. 419-425.
Choi, J.H., Lee, B. and Lee, S.H. (2023) ‘The Role of Generative AI in Enhancing Learning Outcomes in Writing Education’, Education and Information Technologies, 28(12), pp. 16569-16588.
Draper, M.J. and Newton, P.M. (2023) ‘A legal perspective on the educational implications of AI-enabled plagiarism: a case-study of the United Kingdom’, International Journal for Educational Integrity, 19(1), p. 11.
Eaton, L. (2023) ‘How ChatGPT and other AI text-generators could revolutionize academia’, University Affairs, 11 January. Available at: https://www.universityaffairs.ca/features/feature-article/how-chatgpt-and-other-ai-text-generators-could-revolutionize-academia/ (Accessed: 15 May 2024).
Kerr, I. and Rilat, E. (2023) ‘Hey, ChatGPT! What Should We Teach Law Students About Generative AI?’, Social Science Research Network. Available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4481358 (Accessed: 15 May 2024).
Mata v. Avianca, Inc., No. 22-CV-1461 (PKC), (S.D.N.Y. June 22, 2023).
Perlman, A. (2023) ‘The Implications of ChatGPT for Legal Services and Society’, Harvard Journal of Law & Technology. Available at: https://jolt.law.harvard.edu/digest/the-implications-of-chatgpt-for-legal-services-and-society (Accessed: 15 May 2024).
QAA (2023a) Generative Artificial Intelligence in assessments: An academic integrity and assessment security perspective. Quality Assurance Agency for Higher Education. Available at: https://www.qaa.ac.uk/docs/qaa/news/generative-artificial-intelligence-in-assessments.pdf (Accessed: 15 May 2024).
QAA (2023b) ChatGPT and Artificial Intelligence: Q&A for HE providers. Quality Assurance Agency for Higher Education. Available at: https://www.qaa.ac.uk/news-events/news/chatgpt-and-artificial-intelligence-q-a-for-he-providers (Accessed: 15 May 2024).
Salmon, G. (2023) How to design assessments in the new world of AI. Wonkhe. Available at: https://wonkhe.com/blogs/how-to-design-assessments-in-the-new-world-of-ai/ (Accessed: 15 May 2024).
SRA (2019) SRA Principles. Solicitors Regulation Authority. Available at: https://www.sra.org.uk/solicitors/standards-regulations/principles/ (Accessed: 15 May 2024).
Susskind, R. (2023) Tomorrow’s Lawyers: An Introduction to Your Future. 3rd edn. Oxford: Oxford University Press.
