The four governance components depicted in the previous post can have highly diverse configurations that depend to a large extent on the specific social characteristics of the groups harnessing to reach given outcomes. Given our core topic, AI governance, my main focus here at the macro level, that is, on specific socio-economic formations, particularly on modern states that have, since the beginning of the so-called Modern Era, shaped the political sphere based on given power distribution among contending groups and classes. Modern states exercise power by designing and implementing public policies that distribute economic, social and political benefits to the various groups. Such a process is contentious for the most part.
It is critical to distinguish between governance and policy design and implementation. Public policies set the outcomes society wants to achieve in the medium and long term. The social ship then needs to be steered in such a direction. Policymaking thus entails some form of governance, that is, a set of structures and processes that foster its design and implementation. They can range from simple arrangements, as is the case with monarchies and other forms of authoritarian rule, to complex ones, as is the case of advanced capitalist countries where a vast set of public institutions exist, deploying diverse governance mechanisms to support decision-making processes.
Ensuring that public policies benefit the whole of society is critical. Governance plays a twofold role here. First, it is a means to reach the agreed-upon outcomes. Second, it is also an end that ensures the outcomes sought are indeed the result of governance processes and structures that factor in the voices of all those directly and indirectly affected by the proposed policies. Governance processes that are open, inclusive, participatory, transparent and responsive are best suited to achieve such a goal. In this case, we can speak of “good” or “democratic” governance. Note, however, that so-called democratic regimes do not consistently implement such a practice. For the most part, public institutions have not been designed to enable, on a binding basis, such “democratic” governance processes.
The state’s history in the Modern Era yields an interesting pattern. Feudal states in the West and kingdoms elsewhere around the globe were able to deploy state mechanisms to collect taxes from peasants et al. That led to state posts for tax collection and revenue administration. These posts were usually filled at the monarch’s discretion or sold to the best bidder, with qualifications never being a deciding factor. The adoption of the rule of law to legitimize tax collection came next, often based on Roman Law in the case of the Absolutist State in the West. From a governance perspective, we thus start to see the introduction of new structures and processes deemed necessary to sustain the monarchy, but typically all under the direct oversight of the King or Queen.
The French Revolution opened the door to and legitimized the participation of the Third and Fourth Estates in political processes and placed political representation at the center. However, that did not immediately translate into direct involvement in policymaking processes. The consolidation of capitalism in the 19th Century found its ideal partner in the nation-state, which then became the ideal type for state-building around the globe. Hegel’s Philosophy of Right is perhaps a much-required reading here, not only Hobbes, as we are usually told. Nation-states developed a series of complex public institutions, starting with Central Banks and sectoral ministries catering to specific state functions and people’s needs and demands.
In the West, nation-state competition led to wars in the first part of the 20th Century, which brought utter destruction, with the atomic bomb being its undisputable sovereign. Reconstruction and recovery led to the golden age of capitalism, supported by almost universal Keynesian policies, that ended almost abruptly in the 1970s amid a deep economic crisis. The emergence of Neoliberalism as the new socio-economic and political paradigm questioned the capacity of the state alone to manage and regulate markets and shifted the onus to the private sector. The governance structures and mechanisms of an excessively centralized and working-solo state were questioned.
Not surprisingly, public administration perspectives such as New Public Management became fashionable, copying the inner workings of the business sector, which was now calling most of the shots within the public sector. Clearly, the participation of non-state actors in public governance structures and processes started back then. The emergence of the Internet and the widespread use of digital technologies facilitated the involvement of additional state actors in public governance processes. This ranged from typical stakeholders to CSOs and other non-profit organizations representing diverse actors and themes.
The above schematic state history thus reveals its dynamic nature. Although moving at a languid pace, we can conclude that the state has come a long way regarding governance processes. However, that does not mean that existing governance mechanisms in public institutions have changed dramatically since then.
The involvement of non-state actors in governance has become more common and prominent nowadays. First, it is not unlikely that stakeholders representing a wide variety of sectors, groups, or classes are involved in designing the societal outcomes targeted by public policies. New governance processes and mechanisms have been deployed to facilitate such evolutions. However, public institutions’ governance structures have not changed dramatically since then. While stakeholder engagement and participation might be welcome, it is only in rare instances that such participation is binding. More often than not, the goodwill of policymakers opens the door here, and even in such cases, inputs provided by non-state actors might end up being totally ignored. In any event, we can conclude that governance processes are changing faster than governance structures. That, in turn, raises the question of how far the state can change to accommodate the new participatory environment, assuming the nation-state is here to stay. Alternatively, we could also think of other options that could work locally and globally. However, such critical issues go well beyond the scope of this blog.
At any rate, that is the context in which we should frame the governance of AI. For starters, it should now be evident that AI governance is not the outcome we are trying to pursue. Here, governance should be seen as a means to achieve a set of goals. Some could argue that “responsible” AI might be one of them. Others would like to link it to broader socio-economic and political outcomes, especially in the Global South, where many lagging countries still face dire conditions. Having AI governance alone there will certainly not make a substantial difference.
Furthermore, AI governance faces issues similar to many other developmental areas. While new governance processes can indeed be deployed, changing governance structures is more challenging. That is even more acute in nation-states with little state capacity, which cannot successfully deliver most of the public goods and services under their purview. Public institutions there are underfunded and lack sufficient human capacity to initiate unprecedented structural changes. That is partly why AI governance is so relevant to most countries.
Finally, let us not forget that AI, both numeric and generative, has the potential to impact AI governance and overall policymaking processes in ways not previously possible. AI in governance thus becomes a new topic that needs to be framed within governance structure and processes.
Raúl