GENERATIVE Artificial Intelligence (AI) presents a myriad of opportunities for integrity actors – anti-corruption agencies, supreme audit institutions, internal audit bodies and others – to enhance the impact of their work, particularly through the use of large models (LM).
Language models are an evitable part of speech enabled applications that require converting speech to text and vice versa. As part of conversational AI systems, language models can provide relevant text responses to inputs.
In light of this, AI-anti corruption technologies are clearly important because integrity actors like anti-corruption and internal audit bodies are very important towards minimising, rooting out and controlling corruption or unethical activities in the private and public sectors of the economy. As a preventive tool, AI anti-corruption technologies can predict corruption patterns with accuracy.
For example scholars have used a neural network approach to develop an early warning system to predict corruption using political and economic factors (e.g. economic growth, political party endurance).
Neural network approach is critical because it is a method in artificial intelligence that teaches computers to process data in a way that is inspired by the human brain.
In essence, it is a type of Machine Learning (ML) process, called deep learning that uses interconnected nodes or neurons in a layered structure that resembles the human brain.
Keep Reading
- Time running out for SA-based Zimbos
- Sally Mugabe renal unit disappears
- Epworth eyes town status
- Commodity price boom buoys GB
AI anti-corruption technologies (AI-ACTs) can be considered as a socio-technical assemblages based on the use of AI techniques; their scope is to address corruption, while their more immediate aim is to address various types of corruption and related behaviours, ranging from petty corruption to grant corruption(Mattoni et al……).
Seen this way, the structure of an innovative network should be conceptualised as a broad socio-technical assemblage, something that encompasses ongoing interactions among human and non-human actors (de la Bellacasa, 2011; Licoppe, 2010, Mcloughlin et al, 2000).
More so, AI can really aid Southern African integrity actors in combatting corrupt activities that are ubiquitous in every segment of their economies.
The corruption range from financial embezzlement, misappropriation of land, tenderprenuership, smuggling of goods at immigration centres, gender-based violence, murder, robbery, violent conflict and so on.
In some instances, corruption causes serious leakages to the economy, resulting in poor service delivery hence causing untold suffering to the majority of the citizens.
AI–ACTs allow for the fast processing of large volumes of data and thus represents a game-changer in anti-corruption: Thanks to their features, these technologies can promote transparency, reduce the discretion of public officials, and mediate government-citizen interactions, Caroline Gerli PhD Research (AY2023/24). Processing of large volumes of data affects most agencies of government including those of the private sector.
This generally impacts negatively against the processing of corruption cases by the Zimbabwe Anti-Corruption Commission or any other integrity actors in Southern Africa.
AI uses several methods to process data, including data preprocessing: AI can clean normalise, and convert data. Missing values, outliers, and inconsistencies are addressed to prepare raw data for analysis.
Data transformation: AI can convert data formats to work with certain analytic tools and models.
This is explained by AI data processing features that include automated machine learning(AutoML) for analysing text data, predictive modelling, real time analytics, and visualisation capabilities.
These features enable users to manage data analytics workflows from data preparation to visualisation with minimal coding.
With this kind of technology definitely corruption cases processing will be expedited by the anti-corruption bodies and integrity actors in Africa and Southern Africa, in particular.
The most crucial aspect is the need for African governments to plough a lot of resources in the AI facilities that have the potential to process data, especially data meant to combat corruption or graft activities, which is costing most of their economies huge sums of money.
The other key aspect is ensuring that the personnel that work at the anti-corruption commissions or integrity actors that prevent corruption in various Southern African countries are proficient in using AI data processing equipment. The capability to use AI hardware by these officers will ensure that they expeditiously process corruption related data.
This can then arm these corruption prevention agencies with the evidence to report the matter to law enforcement agencies hence trial and conviction would be effected to recover the misappropriated resources, if necessary.
Southern African central governments and their legislatures need to come up with policies that encourage research, innovation and technologies that can avail tools for detecting corrupt activities in public procurement, embezzlement, graft, and anti-competitive practices, by analysing tenders, bid submissions and contracts.
Notable examples, are those in Brazil where the World Bank piloted the Governance Risk Assessment System(GRAS) tool launched in November 2023 that uses advanced data analytics to improve the detection of fraud, corruption, and collusion risks in government contracting. In China, public authorities developed the "’Zero Trust system, in which ML was used to predict the risk of public officials engaging in corrupt practices ;unfortunately ,the system was allegedly abandoned for being too efficient(Chen,2019). Be that as it may, Southern African policymakers need to appreciate that there are also challenges associated with using AI in anti- corruption and integrity organisations.
One of them is inter-disciplinary challenges that can result in underestimation. One very important aspect is the procurement, and development of AI tools hence efficient policies and regulations are strictly needed. AI tools to curb corruption can suit certain contexts, in other words the way it is designed and implemented would normally mirror a certain society.
Furthermore, not only can the design, development, and implementation of AI–ACTs be corrupted itself (especially in autocracies), but the need for these technologies in public organisations also constitutes a profitable business opportunity for Information Technology(IT) private sector actors. Southern African scholars/academics need to make serious inquiries into the use of AI tools to combat graft in private and public sectors of the economy. These research interventions are key towards minimising corrupt deals that are detrimental to the wellbeing of citizens. It is a fact that leakages in the economy deprive citizens of their national cake in full.
Mostly this deprivation through corrupt tendencies by persons in public and private authority manifests itself in poor service delivery by both local and central governments. Generally speaking, scholars agree that AI can be a "positive shock’’ (Colonnelli et al., 2020) in the monitoring capacity of public organisations, especially concerning risks assessment and third-party due to diligence.
However, a broad consensus exists on the auxiliary role of AI in the fight against corruption: AI should complement existing anti-corruption efforts rather than replace them entirely. As Etzioni and Etzioni (2017) underline, AI systems should be ’partners’ rather than ‘minds’. Furthermore, scholars have argued that a critical success factor of AI–ACTs is the overall ecosystem in which they are designed, implemented, defined by data input, design of algorithms, and political and socio-cultural institutions in place.
The current challenges in Southern Africa are that the AI meant to detect crime are predominantly made from the West with no or very little input from Africa or Southern Africa itself. Hence the need for researchers in Southern Africa to come up with innovative solutions towards the designing, development and deployment of AI tools that responds to Southern African corruption trends, regarding collecting and collating data.
If this is achieved, it can assist in the reduction of corruption in all its forms and shades. The researches may need to design, develop, and implement AI tools that can detect the drivers of corruption. African and Southern African researchers would need to learn from countries that have long designed, developed and implemented these AI tools for combating graft.
So it may be key that collaborative synergies are created between north and south-Universities, polytechnics and independent research hubs. These collaborative approaches will assist African researchers to understand the concept of Generative Artificial Intelligence and AI tools for combating graft in particular. This will also help them to come up with AI anti-corruption technologies that suit the local environment in combating crime.
- Mabhachi is a freelance journalist and wireless technologies and dynamic spectrum access activist. He is contacted on mediatechzim@gmail.com.