The three propositions explored here for Scoping Cybernomos provide the theoretical spine with which to frame it as an institution; the discursive intention to be necessarily ambitious, prescriptively provocative. They were first conceived to address the problem of alignment. A response that could be characterized as something of a back-to-the-future discursive maneuver, in that all three formulations were either directly or indirectly influenced by monotheism. Subsequently understood here as a sort of agnostic theology for governance in the age of AI.

The Sanctity of Life

It would be reasonable to suggest that the influence of monotheism would have gained more traction if backdated into our past by some 70 years or so; hence the rhetorical interest in courting that possibility. Nevertheless, the idea that religious authorities might have funded our putative Cybernomos Institute back then, does not preclude them from doing so now. In fact the institutional value of engaging with religious authorities is fundamentally congruent with the emphasis on the sanctity of Life endorsed here. Once again it is the enormity of the task that needs to be registered; the comprehensive recalibration of horizons that needs grasping. Arguably it is primarily the cultural institutions of faith that necessarily carry the moral weight; that can absorb the importance of this technological development; that can address the concentrations of power; that have rightly put the sanctity of Life as one of their preeminent goals.

The Precautionary Principle

What’s of focus is how governance structures need to comprehensively adapt to the pervasive influence of machine intelligence. When developing the proposed agenda the rationale behind the precautionary principle proved central. This principle is understood as a broad philosophical and legal framework. One that dictates that if an action or policy has a suspected risk of causing severe harm—to the public, the environment, or a system—the absence of full scientific certainty should not be used as a reason for postponing cost-effective measures to prevent that harm. Because the AI alignment predicament is arguably an order of magnitude more important than any other issue facing the future prospects of humanity, the phrase ‘cost-effective’, whilst understandably grounded in lived economic reality, is actually meaningless. Nonsensical in that there is no future if this cost is not met, whatever that may eventually be. That is you can credibly couch this in terms of existential risk, along with perhaps more immediate concerns, politically, with unwarranted concentrations of power that threaten the meta-governance achievements of representational democracy.

Three Propositions

The propositions are in order of their conception. A triad of interrelated research directions within the proposed Cybernomos Institute. The intention is to provide a scaffold, a framework, within which specific policies can evolve. The idea, the governance objective, is that it has real universal application. That could theoretically be adopted by nation states, that might have existing antithetical governmental obligations. In this scoping exercise such prescriptive modeling should be regarded as a sincere attempt to outline a program, not to suggest the final product. There will no doubt have to be a suite of cultural developments to effectively manage the wholesale adoption of machine intelligence. The virtue of scoping cybernomos is that we already have significant cultural resources in place to address this issue, that are able to embrace its necessary ambition; that can deliver a coherent response to the demands of the precautionary principle. The three propositions are:

[1] TOKII—the Tree of Knowledge Irony Indicator. What should the communication approach be regarding the radical indeterminacy of AI? How do you develop a communication and educational strategy that encompasses both the promise and risk associated with machine intelligence.

[2] ASI—the principle of Altruistic Self-Interest. Given that emergent AI is set to be a meta-governance event in human history, this will also arguably require, at the intersection of AI and governance, an equally profound adjustment and extension of our moral horizons. What should such moral principles look like? Is the ambition of ASI a good candidate for framing that response?

[3] SAI—the goal clarifying notion of Symbiotic Artificial Intelligence. By inviting a computational intelligence to share the cognitive world we humans live in, what should or could such a relationship entail? How much of the governance of the world are we willing to seed to the control of computational intelligence, and how much do we need to? How much adjustment do we need to make if we are serious about evolving our moral horizons to meet the challenges of AI? What actually constitutes effective symbiosis with AI, and are we willing to evolve with it?

These three directions are conceived to be potential beneficial prospects toward the governance of AI, or, as in the first of these—TOKII, a means of productively thinking about the framing, and the communication, of the governance response.