Page 66 - UKZN Proceedings of the Conference Report

There are recommended measures to optimise the benefits of AI while overcoming these difficulties. One way is to create language-based AI which is localised for regions. This would facilitate the use of AI by many people in the population, as suggested by Beniwal and Singh (2024). Equally, protection of the free-market principles by the State and its institution will aid the private sector in raising funds for the development of AI and its application in society. Furthermore, governments may also adopt measures to influence fiscal policy to support their citizens against the negative economic effects of AI. As Beraja et al. (2023) observe, administrative and political transformations geared toward mitigating the social effects of AI, especially job loss, must be carried out by regimes so that they may stay in power and receive public support. Celeste (2022) explains this by portraying how Brazil’s Internet Bill of Rights provided for the same regional technology development while accommodating different actors through a collaborative governance model. Ovadya (2023) is of the view that policies will be more successful in the age of AI if democratic processes are enhanced through citizens’ assemblies and open decisions.
Balancing the potential benefits of employing AI in the public sphere, creating new risks and responsibilities also comes to the fore. The influence of AI in respective sectors is on the rise, and with it the challenges related to disruption of labour markets, privacy, security, and social inclusion (OECD 2019a; World Economic Forum 2018). Among these worries, those related to ethics, fairness, bias and inclusion are most acute, according to Yudkowsky (2008). The way AI works and decides not only promotes the risk of such failure but raises the risk of the lack of accountability itself in this situation, when decisions that are made are based on algorithm- driven judgement (Wirtz et al. 2020). According to the OECD (2022) member countries should be able to establish a regulatory regime that is clear, exhaustive and classifies technologies progressively as they evolve. As Fountain (2019), Cath et al. (2018) and Bryson and Winfield (2017) observe, there are considerable advantages of AI technology in solving such ‘wicked issues’ as political violence or threats to society, but one has control over this technology for its own good.
Notwithstanding the inherent challenges of implementing AI systems, there is great enthusiasm in most countries as they believe that such tools can make public sector governance more efficient and stable. To illustrate, one of the aspects of government operations that can be automated using AI is administrative processes, thereby enhancing government productivity levels and freeing up humans from some simple processes to focus on more difficult ones (Ojo et al. 2019). Al-Mushayt (2019) remarks that employing AI for e-government services can facilitate their effectiveness and responsiveness to the
needs of citizens. Further, the importance of AI functions is on the rise, particularly in risk identification and urban management. In this regard, AI can be applied to urban surveillance to enhance systems of fraud prevention and detection as well as to support the combatting of crime (Ojo et al. 2019; Bullock 2019; Gomes de Sousa et al. 2019). The boom in Smart Cities brings additional chances for AI to boost public administration by designing infrastructure and improving the standard of living for residents (Ben Rjab and Mellouli 2019).
The concept of public governance put forth by Fukuyama (2013) refers to the process of rulemaking and service provision to citizens. The aim of governance is, according to Lynn et al. (2000), to produce as well as to render goods and services within the frames of existing laws, which anticipate the cooperation as well as decision-making of actors from the government sector, private market and civil society. As mentioned by Wirtz et al. (2020), the application of AI in this space is burdensome, as governance brings more risks, such as concerns related to privacy, safety, or even accountability. AI also provides a chance to make government more responsive, assist in better management of cities, as well as serve citizens better.
As AI capabilities keep improving, there is a need for policymakers to figure out how best to place it in the existing governance systems. The introduction of advanced technologies and changing paradigms in social organisation are often much quicker than the response of public administration, with the result that there is a deficiency in governance and legislative (Wirtz et al. 2020). Wang and Siau (2018) point out that the global governance framework must not shy away from the risks posed by AI but build on the strengths of the emerging technologies. The main challenges to successful incorporation of AI into public administration lie in creating public confidence in the deployment of AI while at the same time going ahead with regulation, ensuring there is always control over the deployment of AI in governance for the good of society.
Environmental conservation and healthcare
There is increased interest globally in improving AI for the purposes of environmental conservation. To illustrate, Oliveira and Siqueira (2022) note that AI is capable of estimating the rate of deforestation in the Amazon rainforest by using satellite data, and alerting authorities for action on illegal timber harvesting in a very short period of time. This has been seen to be effective and timely, as opposed to the conventional methods, which mostly relied on people and were slow to enact. AI could provide real-time analysis of satellite images which help in detecting any land cover changes associated with deforestation. Hence, it makes it possible for governments and organisations to take action faster and more accurately,
64 | Proceedings of the conference on Public innovation, develoPment and sustainability