“If we use, to achieve our purposes, a mechanical agency with whose operation we cannot efficiently interfere once we have started it, because the action is so fast and irrevocable that we have not the data to intervene before the action is complete, then we had better be quite sure that the purpose put into the machine is the purpose which we really desire and not merely a colorful imitation of it”[2]

Does Norbert Wiener’s concern within the field of cybernetics effectively function as a precursor to the problems being presented now within the field of computational intelligence. Can his concerns be reasonably understood as an early formulation of the AI alignment problem? Particularly the difficulty of ensuring that a machine’s pursuit of an assigned objective actually serves the human purposes behind it. If you examine the corpus of his writings on the subject on what he would describe as autonomous thinking machines it’s reasonable to extrapolate his concerns to those relating to contemporary AI governance. Whilst not specifically addressing what we now understand as AI, conceptually he is still in the terrain of autonomous machines processing information and speculating as to how they can be controlled. Especially the implications if the technological construct is more intelligent than the inventor, more productive, and faster to output when compared to a human counterpart.

“It is my thesis that machines can and do transcend some of the limitations of their designers, and that in doing so they may be both effective and dangerous”.[3]

What Wiener’s prescient calculated concern arguably adds up to is the first draft of two interconnected discursive positions endorsed here. Namely that the advent of AI is a Meta-Governance event in human history, and that given that significant, rare nomination, it now requires an equally seismic response from the evolving moral status of humanity. That is can Wiener’s reaction to the technological advances of his time, to the possibilities and dangers, be interpreted as an apprehension this would evolve into a meta-governance event in human history, the like of which we now arguably have with AI. The value of his insight, which incorporates a vision of the likely scale of change, would also seem to be accompanied by the necessary ethical considerations that need to be thought through as a consequence of this. His concerns effectively set the stage for what happens over the next seventy odd years. The value of his insight lies in the repeated conjunction of the explicit relationship made between beyond human capacities and the ethical concern that ensues.

“We have modified our environment so radically that we must now modify ourselves in order to exist in this new environment.”[4]

“As engineering technique becomes more and more able to achieve human purposes, it must become more and more accustomed to formulate human purposes.”[5]

Wieners observation that increased technological power removes the protection previously afforded by our technical limitations. Where inadequate conceptions of human purpose had previously been less destructive because technical limitations restricted what people could accomplish. Greater power removes that protection. We might reasonably infer from the notion of inadequate conceptions of human purpose—and the competing forms of government that have given expression to that inadequacy—that this implicitly refers to the evolving moral status that informs what we like to refer to as the human condition. The technology morality equation-connection becomes especially significant when placed beside his sporadic ever-present commitments to freedom, equality and goodwill that encompass a conception of the human condition, that can be summed up as judicious benevolence.[6]

The presence of which he insists must supply a standard against which the purpose-exercise of increasing technological power can be judged. His pointed concerns can consequently be understood as reaching beyond whether machines perform their tasks successfully, to whether the social order employing them remains conducive to human development. This is a stronger historical claim than simply saying Wiener recognized a possible danger: his response can be interpreted as apprehending the probable scale of the transformation, together with the ethical reckoning it would demand. At one end of the scale his was a wary sensibility, at the other end it is redolent of Old Testament prophetic visionary insight. It need not imply that Wiener explicitly predicted that such technological determinism would entail a Meta-Governance event in human history, but a lifetime on from when these concerns were first expressed his observations have born out to be remarkably prescient. Norbert Wiener died in 1964, on 18 March, aged 69. It’s not too difficult to envisage that had he not died then, and somehow a cloned version of himself was made available for every university on the planet, we would not be needing this scoping agenda, our Cybernomos Institutes would already be in place, and in a much better position to address-assess the risk reward dilemma that is AI alignment.