Picture source: Getty/iStock
The rapid advancements of Artificial Intelligence (AI) have resulted in multiple innovative ways in which AI can be used to help people. This article identifies and explores how recent developments in AI can be used to assist South Africa (SA) in promoting access to justice. This identification is achieved through an exploration of United States and Australian use cases on the adoption of AI to promote access to justice. Lastly, this article discusses key considerations such as cybersecurity and regulation of AI in SA, both of which are aspects, to be considered prior to implementing AI in SA.
In the Department of Justice and Constitutional Development’s Annual Report of 2022/2023, access to justice is defined as improving the protection of rights and making it easier for vulnerable individuals to access the justice system. The report noted information technology plays a critical role in maximising productivity and increasing the efficiency of access to justice (Department of Justice and Constitutional Development ‘Annual Report 2022/2023’ at 51 (www.justice.gov.za, accessed 10-2-2024)).
The Department further seeks to implement technological advances aimed at assisting in the delivery and accessibility of the justice system. These advancements relate to the Master’s office. With the help of digitisation, the Master’s office has introduced self-service computers and is developing an online system for registering deceased estates and trusts (Department of Justice Annual Report 2022/2023 at 51). Furthermore, recent technological developments in Generative AI (GenAI) may also assist in promoting access to justice but firstly, what is AI and GenAI?
AI was traditionally used for pattern identification, classification, and prediction. As a specialised area of computer science, AI focuses on the development of machines that can perform tasks that require human intelligence. Unlike conventional software, AI can learn and enhance its performance over time (Johan Steyn ‘Generative AI will transform the legal profession’ (www.businesslive.co.za, accessed 17-8-2024)).
AI is further subdivided into Predictive AI and GenAI. Unlike GenAI, Predictive AI does not create original content. Predictive AI can complete tasks that ‘do not require emotional intelligence or creativity’. ‘Predictive AI relies on historical data to predict outcomes’ to complete certain tasks (Chris Owen and Mary-Frances Murphy ‘Virtual justice? Exploring AI’s impact on legal accessibility’ (www.nortonrosefulbright.com, accessed 23-3-2024)).
Conversely, unlike previous machine learning models, GenAI can produce new output from patterns it has learnt (Steyn (op cit) at 10). This content is often indistinguishable from output generated by humans (OECD Artificial Intelligence Papers no 9 ‘Generative artificial intelligence in finance’ (Paris: OECD Publishing 2023) at page 6 (www.oecd-ilibrary.org, accessed 4-1-2024)). GenAI makes use of large language models (LLMs) to generate the said output. Furthermore, LLMs can be trained on any content and used to produce natural language text in a variety of styles. LLMs can predict words that are likely to follow from a particular request.
An LLM can be trained on any information available on the Internet. However, to avoid any potential copyright infringements, it would be best to first obtain the necessary consent prior to training an LLM on content that is copyright protected. However, in consideration of how AI may be used in the promotion of access to justice, an LLM can be trained on information in the public domain specifically made available for the promotion of access to justice.
GenAI is an excellent tool for summarising and translating documents described as ‘legalese’ into simple and user-friendly forms (Quinten Steenhuis ‘Here’s how AI and ChatGPT tools can help scale pro bono work’ (https://news.bloomberglaw.com, accessed 5-1-2024)). Furthermore, Stanford Legal Design Lab and Suffolk LIT Lab, in the United States, have built an application that assists Suffolk clients with pro bono matters. This application takes the form of a game and is called ‘Learned Hands’.
Standford Legal Design Lab invites lawyers, law students and other legal professionals to play the game to assist it in building LLMs. These LLMs are used to connect people to high-quality legal assistance, legal resources that are jurisdictionally correct, issue specific, and immediately actionable. By spotting trends in the questions people pose, Learned Hands can provide direct services with a more general understanding of ‘legal health’ trends (https://betterinternet.law.stanford.edu).
Learned Hands is used to do basic research into key questions around access to justice. Furthermore, it identifies clients’ key legal needs, how they are clustered or patterned together and how they are expressed. This assists in automatically identifying high level and specific legal issues in clients’ stories.
The game is mobile-friendly and web applicable, wherein, lawyers and law students can earn pro bono credit for playing the game. Similarly, s 29(1) of the Legal Practice Act 28 of 2014, makes for provision for the Minister to prescribe the requirements for community service to be conducted by candidate legal practitioners and legal practitioners. Thus, a similar application of Learned Hands in SA could assist in promoting community services by candidate legal practitioners and legal practitioners.
Similar to the Learned Hands is the ‘Justice Connect’ program. This is a not-for-profit law program, based in Australia, which commenced in 2017. The aim of the project was to scale its impact through technology driven products and processes to support its work in increasing access to justice. In 2021, Justice Connect piloted and launched a Pro Bono Portal, which connected Community Legal Centres and people with unmet legal needs. Furthermore, following the success of the Pro Bono Portal, it has now been deployed in Hong Kong, Germany, and the Philippines (Justice Connect ‘Our history of innovation’ (https://justiceconnect.org.au, accessed 29-3-2024)).
In 2022, Justice Connect underwent an AI intake integration into its tool to assist people. With the help of AI, Justice Connect assists people in identifying their legal problems in real-time while gathering a collection of language samples from diverse and marginalised communities. These language samples are used to train its model to work consistently across various demographics pursuant to assisting various kinds of individuals (Justice Connect (op cit)). Similarly, in SA, a country rich in diversity, an AI intake integration aimed at assisting people from all kinds of backgrounds could present an opportunity in the promotion of access to justice. Perhaps even making the law available in different languages.
Both the Learned Hands game and Justice Connect indicate the potential ways in which GenAI can be used within SA to promote access to justice. With the use of LLMs being constantly trained by legal practitioners; South Africans will have access to high end legal assistance. This assistance may take the form of a conversational dialogue that is jurisdictional specific, issue specific and immediately actionable. Organisations can now spot trends based on what clients ask and provide direct services with a more general understanding of legal health trends. However, notwithstanding that AI presents major opportunities the implementation thereof, also presents some challenges, namely, in relation to regulation of AI in SA and cybersecurity.
Cybersecurity seeks to protect systems, networks, and programs from various digital attacks. These digital attacks may take the form of a cyberattack, which is aimed at accessing, changing, or destroying sensitive information (Cisco ‘What is cybersecurity?’ (www.cisco.com, accessed 20-8-2024)). The same was evident in the Companies and Intellectual Property Commission’s (CIPC) recent attempted security breach resulting in the compromise of both CIPC employees and clients’ personal information.
Furthermore, in 2023 the Information Regulator issued a R 5 million administrative fine against the Department of Justice and Constitutional Development for the contravention of Protection of Personal Information Act 4 of 2013. As mentioned in a statement released by the Information Regulator, the Department was required to submit proof to the Information Regulator that the Trend Anti-Virus licence, the SIEM licence and the Intrusion Detection System licence have been renewed. These licences are imperative in safeguarding the security of the Department (Information Regulator ‘Media statement: Infringement notice and R 5 million administrative fine issued to the Department of Justice and Constitutional Development for contravention of POPIA’ (https://inforegulator.org.za, accessed 9-3-2024)). Therefore, should the Department seek to implement AI developments, cybersecurity is an imperative component in ensuring South African citizens’ information is securely processed.
Secondly, SA currently does not have any legislation dealing specifically with AI. However, the G20 Ministerial Statement on Trade and Digital Economy has set out certain guidelines on the incorporation of AI by a particular member state. One note refers to human-centred AI. This approach envisions inclusive economic growth, which will be of great benefit to society as it aims to empower individuals (G20 Ministerial Statement on Trade and Digital Economy at para 17 (www.mofa.go.jp, accessed 20-8-2024)). An inclusive adoption of AI is imperative, especially, within the context of using AI aimed at the promotion of access to justice.
Furthermore, the European Union has adopted a ‘rights’ driven approached in relation to the regulation of AI. This approach focuses on the intervention of policy makers to ensure citizens’ rights are not breached for AI innovation. Conversely, the United Kingdom and United State has taken a market-based approach (Preeta Bhagattjee ‘International: Trends in AI governance’ (www.dataguidance.com, 6-3-2024)). With a market-based approach, the focus is on promoting technological progress and innovation.
Notwithstanding the aforementioned, on 5 April 2024, the Minister of Communications and Digital Technologies hosted a National AI Government Summit. This summit launched the process toward the development of a National AI Policy and Regulatory Framework within South Africa. The discussion included a draft document titled the AI National Government Summit Discussion aimed at crafting a South African National AI Plan wherein the government seeks to leverage the advances in AI for human good (Ellipsis ‘Artificial intelligence policy and regulation in South Africa’ (www.ellipsis.co.za, accessed 28-4-2024)).
In conclusion, recent developments in AI present an opportunity for the promotion of access to justice. As seen in the United States and Australia, the implementation of AI can connect people to high-quality legal assistance, legal resources that are jurisdictionally correct, issue-specific, and immediately actionable. However, notwithstanding that AI presents challenges namely from a cybersecurity and regulatory point of view it may present opportunities in the promotion of access to justice.
Marciano Van Der Merwe BA (Industrial Psychology and History) LLB (Wits) is Corporate Counsel in Johannesburg. Mr Van Der Merwe writes in his personal capacity.
This article was first published in De Rebus in 2024 (Oct) DR 18.
