As the legal profession enters the era of generative artificial intelligence (AI), legal professionals must understand both the opportunities and the risks associated with this transformative technology.
This article explores how AI can enhance legal operations and outlines the ethical and practical considerations necessary for its responsible use. The first aspect to consider is legal operations, and the second aspect is legal risks and how to mitigate them.
AI in legal operations
Legal practice is evolving, and it is important to consider the potential that lies within generative artificial intelligence to enhance operational efficiency, decision-making, accuracy and strategic capabilities. Lawyers who effectively leverage artificial intelligence will gain a significant advantage over those who do not.
Generative AI is both a practical and immediately applicable tool for legal professionals. Generative AI can optimise all aspects of a legal professional's daily work, from preparing meeting agendas, reviewing e-mail chains, assigning action items, preparing meeting minutes, to the more nuanced areas of legal research, contract reviews, contract management, preparation for litigation, conducting due diligence processes and assisting in regulatory audits.
It is strongly encouraged that all legal professionals start adopting and experimenting with this new technology to improve their AI literacy. Simultaneously, it is important to keep abreast of international regulations on the use of AI, such as the European Union's comprehensive, risk-based AI Act.
The challenges of AI in the legal industry
AI integration within the legal space is not without challenges. Key considerations involve data privacy, confidentiality and intellectual property rights. Legal practitioners must consider security measures and develop AI usage guidelines for their employees' use of AI. Most legal practices will adopt off-the-shelf generative AI tools such as the general-purpose Microsoft Copilot or Google's Gemini.
Safeguarding confidential information
We recommend adopting enterprise versions of AI to safeguard confidential information, data privacy and observe our obligations as a legal profession, given that enterprise versions do not use input data to train models. Additionally, we advise the adoption of an internal policy that limits employees' use of AI to the firm's licensed enterprise version to avoid the risk of individual employees using their personal, unauthorised, free version of ChatGPT, etc. known as "shadow IT", which exposes the firm to breach of confidential information or data privacy due to the lack of security and privacy safeguards. Given the benefits of AI on day-to-day workflows, it is likely that employees will want to use an AI tool to enable them to be more productive. It is important, then, that enterprise licences are adopted and distributed to employees within law firms to reduce this risk. All employees should be provided with AI literacy training to empower them to use these tools without mistakenly sharing confidential information or causing a data breach.
Understanding the limitations and ethical implications of AI usage
Employees will need to learn prompt engineering skills and understand limitations and ethical implications of AI usage. It is critical that legal professionals must maintain the highest standards of professional judgement and must review all AI outputs for errors. AI requires human oversight given its ability to hallucinate and its tendency to provide biased outcomes. By way of example, in the Colombian case of Espitia Chávez v. Salud Total EPS (2023), a judge used AI to assist in a ruling involving a child with autism. While AI was used for efficiency, the case raised concerns about bias, underrepresentation of minorities, transparency and the influence of non-legal data on judicial decisions.
From this judgment, it is important to ensure that the limitations of AI are always front of mind.
Reviewing outputs
Outputs must be reviewed and verified similarly to reviewing a junior professional's work. Unfortunately, too often legal professionals are not overseeing outputs, and this results in incorrect advice being provided to clients and misleading evidence being presented in court.
In the infamous American case of Mata v. Avianca, Inc. (2023) and the United Kingdom's case of Harber v. HMRC (2023), attorneys were found to have cited non-existent cases following hallucination, misleading the court and prompting a stern reminder to verify case law before quoting it. The South African case of Mavundla v. MEC: Department of Co-Operative Government and Traditional Affairs KwaZulu-Natal and Others [2025] ZAKZPHC 2 resulted in the same findings of inaccurate case law references as a result of plausible-sounding hallucinations. Fictional cases not only waste time and incur unnecessary costs and resources but also pose a serious threat to legal integrity and jurisprudence. It also results in harm to an attorney's professional reputation and could result in attorneys of record being ordered to pay legal costs.
Ethical considerations
Ethical use of AI remains a fundamental aspect; legal professionals must always adopt the principle of oversight or "human-in-the-loop."
How to mitigate the risks of AI
To mitigate the risks and ensure safe AI adoption within your practice, we encourage:
- Enterprise AI tools;
- Implementation of internal AI policies;
- AI literacy programs for employees;
- Review all AI-generated outputs; and
- Monitor regulatory developments.
Adopting AI within your law firm
The adoption of AI usage within your legal practice will have a cost implication. It is advisable that firms ask current vendors what AI solutions are already at the firm's disposal as part of their subscriptions prior to assessing and piloting other solutions available in the market. Build an AI strategy with the technology that is already available to the firm.
Copilot is a powerful tool that can be leveraged by legal professionals, and it is a great starting point if the firm is already using Microsoft's Suite. From there, you will be able to assess what additional use cases need to be solved with a developing and maturing AI strategy.
Conclusion
AI holds promise, but its use in law must be grounded in responsibility, transparency, and a deep understanding of its limitations. Legal professionals act as gatekeepers of legal and ethical standards and must act accordingly when adopting this new technology. In doing so, legal professionals will be able to increase their accuracy, research more efficiently and produce better quality outcomes for clients, provided that they adopt ethical practices when using AI and continue to evolve with this technology.
If you're passionate about ethical legal practice, consider enrolling in one of the Gawie le Roux Institute of Law's practical legal training courses. Sign up today to stay ahead of the curve!
People also ask
What is AI literacy?
AI literacy is the ability to understand, evaluate, and effectively use artificial intelligence (AI) technologies and their impact on society. It entails technical knowledge of how AI works as well as critical thinking skills to assess AI's strengths, weaknesses, and various ethical implications.
What is shadow IT?
Shadow IT is the unauthorised use of information technology (IT) resources in the workplace. It includes software, hardware, applications (often cloud-based), and services, without the knowledge and approval of the company's IT department.
What are some of the benefits of AI for lawyers?
Legal professionals can increase their accuracy, research more efficiently and produce better quality outcomes for clients if AI tools are used effectively and ethically.
About the author
Leanne Maroun is the Director of LC Maroun Attorneys Inc., where she specialises in technology and media law. She is passionate about artificial intelligence, legal technology and the tools that can enhance the skills and performance of attorneys.
This blog was last updated on 01/07/2025