According to J.A. Kroll
Work that has long been done by decision-making organizations or individuals is increasingly being done with the help of computerized automation. This shift creates new gaps in governance structures that govern the correctness, fairness and dynamics of important processes and decisions. When existing control systems are not adapted to the speed, scalability and complexity of these new automated systems, any biased, unintended, or incorrect results can go unnoticed or be difficult to correct, even if discovered. These gaps stem from the fact that the nature of artificial intelligence (AI) creates inherent fundamental barriers to governance and accountability in systems that rely on automation. Designing automation in systems is not just a technical issue for engineers and developers. but a question of governance for politicians and users. The management of computerized systems must anticipate and respond to the possibility of creating (or exacerbating) inequalities, and automatic systems must support human values as a means of limiting harm.
To realize the full benefits of automation, governments and technologists must work together to develop governance structures to govern the new capabilities of computer systems. The resulting socio-technical governance structure, or the totality of technical developments, implementation, documentation and guarantees, along with policy, management, maintenance, compensation for damages and enforcement, should include people in the system as active participants, and not just passive recipients of technology. Considering the intended context and technology of an application as a single system, including the people affected and the decision makers who control the structure and function of the system, is the best way to determine which interventions and constraints will reduce the potential harm from its implementation. Building socio-technical control structures requires an understanding of both the systems they govern and the social and organizational context in which they operate. Thus, for proper governance, automated systems must be developed in conjunction with their own governance structures, each of which is designed to support and regulate each other.