The constant evolution and modifications to AI is due to the fact that it is developed by humans for humans. AI affects people in their various social and professional endeavors. And it is only within these contexts that we can adjudge AI to be functional and successful, or ineffective and failed. It is these contextual dynamics of deployment and use that provides the accurate data that enables us to think about the regulatory frameworks that can help us make AI more adaptable for societal, institutional, organisational and ethical use.

The worries about AI are critical ones. There is the worry about the increasing autonomy of AI and the challenges that poses for how humans perceive themselves and their worth. But more immediate is the worry about the societal disruptions of the deployment of AI, especially in the workplace—displacement, alienation, security, death.Thus, AI is not just a panoply of technical issues. It is also fundamentally a human-centered phenomenon that must be adequately understood if we are to better enjoy its functional innovation and creativity.

These issues take on some sinister conceptual and practical directions when situated within Nigeria’s dysfunctional context of public administration. Africa has the unsavory reputation of being the most difficult administrative context in the world. And this translates into a lot of implications and repercussions for individual states like Nigeria. One of the implications is the lack of efficient connection between the public service and the state. It is difficult, for instance, to point to any functional developmental state—in the mold of the Asian Tigers—on the continent. And this speaks volume about the capacity of individual states to deliver on the promises and dividends of good governance for their respective citizens.

It is also a damning indictment on the effectiveness of the trajectories of institutional reforms in Africa. This is the context that demands, as a matter of urgency, the AI revolution in the service of productivity for a citizenry that have been waiting a long time for good governance. And yet, this is where caution is most required in proportion to the level of urgency.

In other words, if the malfunctioning of robot assistants and supervisors can generate such a huge global hoopla within a work and administrative context—like South Korea—that is highly efficient and productive, what would happen if they are deployed within a highly difficult administrative environment? Or, even more fundamental, how do we relate the deployment of AI to a context that less than effective, efficient and productive?

Is it enough to automate, computerise or digitise when the system and processes being improved have not been mapped, reprofiled for reengineering; does it not amount to engrafting technology on a challenged system and what results should we expect? This is the current direction of the institutional and administrative reform dynamics in the Nigerian public service. And it calls for a critical pause for reflection. Two issues are fundamental for resolution. The first is: Can the AI efficiency dynamic be tacked on to a deficient system to achieve efficiency and effectiveness? Yes, it can; but then it becomes another supposedly “innovative” recipe for deepening existing inefficiencies and deficits.

First consideration: we need to start the reflection from the perspective of the self-motivated, hardworking but extremely frazzled and demotivated Nigerian public servant who is compelled by so many factors to work within a highly toxic, inefficient and highly politicised workplace.

This is the first and most significant context that AI is to be deployed. How will this pan out in practice? What regulatory frameworks can such a system deploy? What safety measures can the system afford that will not compromise the human well-being?

Second, how do we ethically mediate the relationship between the robot assistant and the human civil servant not just in terms of emotional connection but fundamentally of ethical relationship. If the existing public service is flawed in mediating human-to-human ethical relations, how do we hope to situate the human-AI component and achieve even a measure of success?

The workplace is a context that must be configured to protect and enhance human dignity, self-worth and welfare which cannot be sacrificed to structural efficiency that AI deployment is meant to address. What accountability structures and standards are then in place to safeguard human self-worth? This also goes beyond the workplace to, for example, the sanctity of data collected by AI. How is the system to ensure data privacy? How about the ethical oversight function of the system to monitor AI autonomy and deployment for critical use? This also affects the way the system manages public perception and public trust with regard to the functional effectiveness of AI.

A challenged system does not need more innovation; rather, it needs a moment to rethink and reengineer and get right the institutional basics. The effectiveness of AI in the workplace is not in doubt; it has been demonstrated all across the world as the harbinger of efficiencies and productivity if properly managed and grafted effectively into a functional system. AI is meant to enhance an already functioning system rather than serving as an instigator for a deficient one.

This implies that to adopt, adapt and deploy AI into the Nigerian administrative workplace must be preceded by an urgent imperative of reflecting on, rethinking and reengineering the administrative and institutional basics that can make the public service genuinely and efficiently worldclass. And the most fundamental question in this regard is: what change management mechanism can yield a government business model that is efficient?
Concluded.

Olaopa is Chairman, Federal Civil Service Commission and Professor of Public Administration.