Chat Generative Pre-Trained Transformer (ChatGPT), a powerful language generation model developed by OpenAI, has garnered significant worldwide attention and utilisation across various industries and professions.

In South Africa, the legal profession is no exception. However, employing ChatGPT in legal practice comes with its own set of risks and dangers. This article sheds light on the potential perils of utilising ChatGPT within the legal sphere in South Africa, emphasising the importance of careful consideration and regulatory measures to mitigate these risks effectively.

Accuracy and reliability issues:
One of the primary concerns regarding the use of ChatGPT in the legal profession is its potential for inaccuracy and unreliability. ChatGPT is trained on historical data and relies heavily on human interaction to learn new facts which generate responses based on the patterns and information it has been trained on, but it lacks real-time judgment and critical evaluation capabilities.

It was recently reported that lawyers arguing in a case in the Johannesburg Regional Court have been called out in a judgment for using fake references generated by ChatGPT. “When it comes to legal research, the efficiency of modern technology still needs to be infused with those of good old fashioned independent reading,” said Magistrate Arvin Chaitram of the situation.

In this matter, although ChatGPT had referred to actual cases and given real citations, the citations related to different cases than the ones named.

Ethical and legal responsibility:
Whilst generative models like ChatGPT provide a great deal of opportunity to the profession, it does come with a large array of ethical and legal concerns which are mostly related to the complex structure of ChatGPT.

Lawyers have a fundamental, ethical, and legal duty to provide accurate and competent advice to their clients.  When utilising ChatGPT, lawyers may inadvertently shift some of this responsibility onto the AI system. If the AI-generated content proves to be incorrect or harmful, the lawyer will bear the professional and legal consequences thereof. 

Lawyers have a duty of technical competence and would be expected to understand how the Chatbot works and what the benefits and risks are if they are to use it in their practice. Lawyers who decide on ChatGPT will also have to be able to intelligibly explain the benefits and risks to their clients.  

Privacy and data security concerns:
ChatGPT operates by processing and analysing large amounts of data to generate responses. Maintaining client confidentiality and protecting sensitive information is paramount in the legal field. At a privacy level, the use of ChatGPT may pose a threat to organisations. Section 1 of the OpenAI Privacy Policy collects information about the user’s browsing activities over time and across different websites. The site does not respond to “do not track” signals.  This may pose a risk to the duty legal practitioners have to their clients. 

The Law Society of South Africa (LSSA) places duties on legal practitioners when using internet-based technologies and in its guidelines, prescribes that legal practitioners are required to take reasonable steps or reasonable protective measures to ensure that information provided by clients remains confidential.

Lack of accountability and transparency:
ChatGPT has been trained on vast quantities of data. It operates as a black-box model. Some of the data included may be proprietary. There can be concerns that ChatGPT may use such material in its response despite not being licensed by the copyright owner. Due to its nature, it’s challenging to ascertain how ChatGPT specifically generates specific responses, and this lack of transparency raises questions about accountability and the ability to contest or challenge AI-generated advice.

In the legal profession transparency and accountability are vital for building trust and ensuring the legal system's integrity.

The main issues surrounding ChatGPT in the legal field include quality control, accuracy, ethical and privacy issues, and the fact that it is not specifically trained in the law. South Africa’s legal community therefore must exercise caution and prudence and do whatever is necessary to mitigate inherent risks associated with the use of ChatGPT.

ChatGPT and similar AI technologies offer exciting potential for revolutionising the legal profession. ChatGPT should however be viewed as a tool to assist legal professionals rather than a complete replacement for human judgment, expertise, and experience.

Careful consideration, ongoing research, regulation, and a clear understanding of the limitations and responsibilities when integrating AI into a legal practice is paramount in balancing the benefits of AI technology with the ethical and professional obligations of the legal profession.


REFERENCE LIST:
1. My Broadband Article “South African Lawyers use ChatGPT to argue a case – Get nailed after fake info” dated 9 July 2023
2. Bloomberg Law Analysis: “Would ChatGPT bring AI to Law Firms? Not anytime soon.” dated 28 December 2022.  By Thomas Bacas 
3. De Rebus Article “Do Legal Practitioners truly understand the danger of ChatGPT?” dated 1 September 2023
4. ENS Article “ChatGPT: A useful tool or risk for Lawyers?” 
5. Cliffe Decker Hofmeyr Article “Should I ask my Lawyer or should I just ChatGPT it?” dated 29 March 2023


WRITTEN BY GERHARD SMIT
Gerhard Smit is a Director, Conveyancer and Notary Public in Miller Bosman Le Roux Attorneys’ Property Law Department.


While every reasonable effort is taken to ensure the accuracy and soundness of the contents of this publication, neither the writers of the articles nor the publisher will bear any responsibility for the consequences of any actions based on information or recommendations contained herein. Our material is for informational purposes.
Powered by SucceedGroup