Page 64 - UKZN Proceedings of the Conference Report

specialists in governance and ethics (World Economic Forum 2023b). These emerging roles show that AI is not simply about enhancing existing industries, but is more disrupting; for instance, prompt engineers responsible for developing AI-based language models are in great demand already. However, other jobs like AI modellers and data trainers are also involved in ensuring that the machines learn to identify patterns and make decisions, which is essential for AI systems to function properly. This also indicates that educational institutions as well as governments will have to look for and support initiatives aimed at upskilling and reskilling individuals to fit into such jobs. In the absence of such education programmes, the workers will not able to operate effectively in an economy where AI is increasingly pervasive.
Although new job opportunities are promised, there are still issues related to the loss of jobs in sectors which have AI technologies that can easily carry out normal work tasks. The threat to jobs is evident in sectors which involve manufacturing, logistics, finance, and most importantly any other industry sectors where repetitive tasks can be performed through automation. For instance, common business applications that have machine learning capabilities and robotic process automation now do simple jobs such as data entry, stock level management, and even basic customer calls, which always previously required human presence. Frey and Osborne (2023) add that the greatest loss or employment shifts due to AI will occur in routine occupations, mainly office and clerical jobs. However, the possibility of losing one’s job is not guaranteed – sectors most likely to be spared the axe of automation are those that deal with art, healthcare and education, where jobs demand creativity, feelings, and tough judgement calls among many other interactions. Additionally, LinkedIn (2023) reinforces the argument that ‘white collar’ sectors like health and education are less affected by AI, as these largely rely on face-to- face interaction and cognitive abilities which machines cannot replicate. Gmyrek et al. (2023) state that while the implementation of AI may increase the workload of 13% of jobs, only 2.3% remains vulnerable to total automation. It can be concluded that even though some low-skilled jobs will be taken over by AI, the overall purpose of such innovation is to boost efficiency among the human workforce and encourage engagement in more productive and innovative spheres. In South Africa there are sectors such as mining, agriculture and hospitality which are at high risk of being automated, but also sectors which may benefit from AI technology through increased efficiency and enhanced working conditions.
The impact of AI goes beyond the effects it has on employment prospects. This is particularly evident in recruitment and human resource management, where AI
is taking a pivotal position. Companies are now utilising AI technology to carry out recruitment tasks, to perform the first round of screening of candidates and, in some extreme cases, candidates are assessed using machines. In 2023 the use of AI for hiring was very high due to the fact that several companies started utilising tools such as Novel, Mya, and Pymetrics, among others aimed at improving the speed of the hiring process. These use natural language processing, predictive analytics, and machine learning to screen CVs, conduct video interviews, and analyse soft skills, thereby reducing the time a company takes to make decisions regarding employment (IBM 2023). This is likely to increase efficiency and fairness in recruitment since AI, in this case, will assist in reducing human attributes and biases in the selection of candidates most fit for the position.
Nonetheless, the application of AI in hiring has also raised ethical issues. As pointed out by Gupta and Mishra (2022), it is important to develop and deploy AI-enabled psychological tools for recruitment in such a way to avoid biases, such as those based on race or gender, or violation of the privacy of the applicants. For instance, if the recruitment process includes use of the internet and social networks to check potential employees, the AI system might be prejudiced in focusing on particular features or actions – which do not contribute to a candidate’s professional fit for the job. According to Radonjić et al. (2022), it is essential to create an ethical framework to recruit with AI without bias, and with transparency and accountability. This highlights the urgency for policymakers and industries to create boundaries on the possible risks of using recruitment systems such as AI.
Besides the short-term effects on jobs, AI is also expected to impact the economy in other ways. According to a study published by PwC (2024), global GDP growth enhanced by AI to the level of up to US $15.7 trillion is predicted by the year 2030. This growth would be mainly driven by productivity levels and very high levels of consumption. It is therefore no wonder that progress in most economies and industries, from commerce and medicine to transport, represents the implementation of AI systems. Nevertheless, economic advantages deriving from AI are unlikely to be achieved equally by everyone. As expressed in the Stanford Artificial Intelligence Index Report (Maslej et al., 2023), the majority of investment in developing AI, as well as innovation activities aimed at it, is already happening in countries like the USA and China, and there are serious limitations in fully exploiting its advantages elsewhere, particularly in Africa. In particular, South Africa has limited access to AI processes, has a poorly developed cyberspace, and lacks skilled personnel. Aghion (2023) also claims that these barriers exist in South Africa, along with other emerging economies, for a reason: to guarantee that AI is utilised to enhance economic growth inclusively. Whether
62 | Proceedings of the conference on Public innovation, develoPment and sustainability