Can AI Govern Society? It’s Not that Simple

Many are beginning to propose A.I. systems as alternatives to govern human society due to problems identified with “democracy;” except, it’s not that simple

First of all, it’s important to remember, as we’ve made clear in other posts, that the systems we currently have are NOT democracies; whatever problems we associate with them, therefore, are not problems or failures of democracies, but problems or failures of autocratic oligarchies (or autocratic “republics”; refer to older posts if needed).

But let’s even assume one still prefers governance by technology to an actual democracy, for some reason. Well:

We must, however, be cautious in our expectations and application of A.I. in governance because not all problems can be solved by data analyses, pattern recognition and/or the reproduction of (however clever) new iterations of common literature or ideas – which is what A.I. tends to be good for, at least today. For instance, in solving the problems we have solved in this book – that is, fixing our systems of governance – we have had to go against popular literature and ideas (which A.I. relies on), and use a kind of logic in analysing theoretical questions or making observations of real world phenomena, that require genuine human intelligence or thinking if you will. It certainly goes beyond that. In solving many development challenges or other human problems, with respect to the formulation of laws or systems, or design of communities with certain sociological considerations, we need a proper understanding, appreciation and, most importantly, experience of human society; and it is easy for those without a background in these areas to sometimes take them for granted, as we might see in a business administrator or technologist’s flawed understanding of governance or development planning, for example, much more in software trained on data, by such a technologist.

Regardless of all this, we can only wait to see where A.I. leads in future. For now, it still has great prospects or applications in the areas aforementioned – such as analysing security threats, economic trends and some kinds of policy and governance decisions…

– The Tragedy Called Democracy in the 21st Century (2023) pp. 357-358¹

What is argued, ultimately, is that, we’re still going to have, or at least need, a human-led form of governance (be it an autocratic oligarchy or, hopefully, a democracy, or other form of government) that can still be facilitated by A.I. rather than to have an actual governance by (or left to) technology; those are two different propositions (a nuance better understood with context provided by an earlier chapter on power and governance, and forms of governance).

Footnotes:

1. Reproduced with permission.