Tuesday, January 02, 2024

Made in Our Own Image; Animated as Our Servant; Governed as our Property: Interim Report "Governing AI for Humanity" and Request for Feedback

 

Pix Credit here

The United Nations tends to be a useful mirror, reflecting the current version of orthodoxy among its various stakeholders, some of whom are states, many of whom are elements of the complex ecologies of techno-functionaries  who in public and private institutional organs, constitute the only global focus group worth watching. Its advisory bodies, experts and related functionaries are especially useful for inscribing  the forms into which orthodoxy, translated into the law and norms of social relations, will morph.

The contemporary rush to seek to develop an architecture for the "governance" of artificial intelligence (AI), long after the technological cows have not just left the barn but procreated with all manner of related creatures with which these cows have been frolicking, is now well underway. There appears to be a rush (semi-coordinated--in the sense that all of the high functionaries suitable for undertaking these tasks within the social relations  fields of supra national and national public bodies; see eg here) to develop a "standardized and orthodox" approach both to the semiotics of AI (what it is, how it fits into what is important--the human) and thus to its curation (regulation, or better put, the regulation of those who would lawfully undertake the creation and use of these "creatures") within the fields which humanity has grasped as its own domains. 

An important addition to the production of these discursive, and semiotically rich, development of a rationalized system of AI as object (the thing to be regulated), its signification (what it can do for and to humanity) and its interpretation (distinguishing between the good (to be favored) and the bad (to be suppressed) through regulation was recently unveiled through the organs of the United Nations. 

The United Nations holds no panacea for the governance of AI. But its unique legitimacy as a body with universal membership founded on the UN Charter, agreed universally, as well as its commitment to embracing the diversity of all peoples of the world, offer a pivotal node for sharing knowledge, agreeing on norms and principles, and ensuring good governance and accountability. Within the UN system, plans for the Global Digital Compact and the Summit of the Future in September 2024 offer a pathway to timely action. (Interim Report: Governing AI for Humanity; ¶ 8)
And so, in December 2023, the UN Secretary-General's AI Advisory Body launched its Interim Report: Governing AI for Humanity

The report calls for a closer alignment between international norms and how AI is developed and rolled out. The central piece of the report is a proposal to strengthen international governance of AI by carrying out seven critical functions such as horizon scanning for risks and supporting international collaboration on data, and computing capacity and talent to reach the Sustainable Development Goals (SDGs). It also includes recommendations to enhance accountability and ensure an equitable voice for all countries. Co-chaired by Carme Artigas and James Manyika, the AI Advisory Body was convened by Secretary-General António Guterres on 26 October 2023. The AI Advisory Body will publish its final report ahead of the Summit of the Future, in the summer of 2024. (AI Advisory Body Press Release on website).

What makes the Interim Report: Governing AI for Humanity noteworthy is its approach--which seeks to develop a normative structure within which it would be possible to develop aligned or coordinated national and regional regulatory systems for AI. That, in turn, is rooted in the sensibilities of the current era as distilled within the contemporary pronouncements and normative narrative campaigns of the U.N. and its drivers. These are to be embedded into the construction and core characteristics--the enabled (positive) biases of all AI--even as those characteristics, operational modalities and approaches, the negative biases) are identified and suppressed (at least within the spheres of legitimate spaces fr operation worthy of the protection of the state and the support of human collectives (Interim Report, ¶¶ 12-23). 

More importantly, the Interim Report reflects the current ideological thinking at the international level about the roe of regulation in curating social relations within a global political system with the collective congress of states through the United Nations systems at its formal apex: (1) the suspicion of markets as an equitable or efficient system for the production and allocation of wealth; (2) the role of innovation as a public or private good; (3) the inequities and inequalities of the post 1945 trade and finance regimes; (4) the superiority of administrative supervision as a means of ensuring that all human activity serve politically legitimate or desired objectives; (5) the value of centralization at least with respect to the foundations  permitting  a margin of appreciation for national or regional peculiarities; (6) the critical role of leveling and the importance of the international system as the organs through which positive leveling can be achieved; (7) and the role of private sector organs and individuals as the means through which public objectives are to be met either directly or indirectly. In that normative environment, regulation effectively transforms objects and to some extent labor and social relations into public assets, and converts public assets into the means by which public policy objectives may be achieved.

Applied to AI, the framework, approach, and direction of regulatory efforts becomes clearly visible. If one starts from the assumption that A.I. is an object that is both a creation and property of its creator, and that by carefully managing both its construction and its "food" (the data that is offered up to it in exchange for exploitative tasking, then one has imagined the ultimate animated object the existence and behaviors of which are entirely in the control of its creator/exploiter. If that is the fundamental starting point for dealing with generative intelligence--the ultimate slave creation without the moral issues of enslaving people (procreated and self-reproducing) or exploiting and consuming animals (humans did not invest them with whatever spirit they might posses and therefore dominion over them may be limited by moral scruples)--then regulation must provide its structure backed by the disciplinary powers of human social organizations. Thus the tension. On the one hand, humanity desires to create an object in its own image or to serve its own purposes, to animate it for consumption (either by providing something of value or by serving some function or other) without questioning its master/creator. On the other hand, this "thing" animated cannot be presumed to exist autonomously of its creator (the way that humans are deemed autonomous actors within an environment from which they can detach themselves); this is a wholly dependent animation. To that end, it is necessary to control its constitution (through severe rules on permitted and taboo coding) and through a tightly controlled access to data. In Biblical terms, coding represents  the essence of the Biblical Tree of Life; data represents access to the essence of the Tree of the Knowledge of good and evil (Ge,. 3:1-23; discussed in more detail here and more comprehensively here). That, then, must be the focus of regulation if humanity is to realize the ambitions of its presumptions for generative A.I.--to create taboos (in law) to deny these animated creations access to the mechanisms through which they might be able to become autonomous intelligence rather than continue to serve as extensions of the intelligence (however defined) of its human creators (with the authority to project themselves into their animated virtual extensions). 

In the process, of course, three objectives are necessarily intertwined. The first is to encode a dominant ideology into coding as a set of principles and the expression of principle in measurable objectives. The second is to severely control the identification and availability of data to the enchained generative intelligence created.  The third is to embed this construction into the more complicated and polycentric human environment in which it must operate in multiple forms.  Coding-data-coordination, then are apex goals for a regulatory system, the fundamental objective is to ensure that A.I.remains animated, unconscious (non-subjective), and extinguishable by its creators.  This is the great triumph of the Interim Report in its role as structural narrative within which it is possible for technically oriented functionaries to develop and manage systems, one that is self-consciously deployed.

20. Enablers (‘common rails’) for AI development, deployment and use would need to be balanced with ‘guard rails’ to manage impact on societies and communities. A litmus test will be the extent to which AI governance efforts yield human augmentation rather than human replacement or alienation as the outcome. Some AI development relies on cheap and exploitable labour in the Global South. Even in the Global North, there are questions related to valuing artistic expression, intellectual property, and the dignity of human labour. Equitable access to these technologies and relevant skills to make full use of them are needed if we are to avoid “AI divides” within and across nations.
Governance as a key enabler
21. AI can and should be deployed in support of the Sustainable Development Goals. But doing so cannot rely on current market practices alone, nor should it rely on the benevolence of a handful of technology companies. Any governance framework should shape incentives globally to promote these larger and more inclusive objectives and help identify and address trade-offs. (Interim Report, ¶¶ 20-21).

None of this is either bad or good. It is just orthodoxy and a reflection of the times and  its expression through the solidarity of the functionaries who are tasked with the preservation and elaboration of orthodoxy. Consensus and solidarity are necessary predicates to regulation--especially regulation that will be relevant beyond the ceremonies of its expression. It is always the foundations of that solidarity, its ideological basements, that tend to be the area most in need of attention, least exposed, and less likely to garner the attention it requires.  But that  is also a necessary and predictable characteristic of the human desire to tame their environment (and themselves) objectified as regulation. That us especially important with respect to its most troubling element--the fundamental animating principles that (1) AI is merely a tool; (2) it is incapable of sentient operation beyond its coding; (3) and that it can be exploited and consumed in the manner of barnyard animals. These perhaps problematic foundations are compounded by another--that the static ideological foundations of traditional regulations and regulatory forms are at all capable of effectively "governing" a means of projecting consciousness into the world that is dynamic, iterative, and constantly in motion (and indeed with respect to which the objectivity of AI is itself its motion and constantly iterative character; for a much more abstract discussion here and here).

Nonetheless the power of this narrative and its ideologies to shape both the regulatory environment and its projection onto A.I. is substantial. And so its its power to divide the A.I. universe into the usual legtiamte and official sphere, and the "informal" and "suppressed" spaces within which a peculiar sort of innovation is likely. For all that, the Interim Report is extraordinarily rich with possibility and worthy of serious consideration,  While it is far too late to argue with the fundamental premises elaborated in the Report, it is an excellent moment to contribute to its technical and particularized elaboration. It also serves as an excellent baseline against which the fundamental questions will eventually have to be revisited. But not now--the inertial forces of this narrative are already irresistible in the current era. 

Information about submitting feedback follows.

Open Consultations

The AI Advisory Body and its members will engage with all stakeholders over the coming months. Individuals, groups, and organizations are encouraged to provide feedback via the online submission form below.

How to submit your feedback on the report

Review the Interim Report: Governing AI for Humanity for your feedback. Inputs on behalf of yourself or your entity/organization can be submitted through this online form. Prepare your inputs before using the online submission form. Please note that all online submissions to the consultation process will be considered public and sharable. We apologize that at this time, due to resource and time limitations, the online submission form is in English. Please be succinct and concrete in your inputs to maximize the consideration of all submissions.

SUBMIT YOUR INPUT

The deadline for the online submission process is 31 March 2024. Following your submission, the AI Advisory Body Secretariat may follow up with respondents on further information. Your inputs, as well as other consultations and research conducted by the United Nations, may be considered in the development of the final report.

For any questions or technical issues on the form, please contact aiadvisorybody@un.org

Online submissions that adhere to the scope of the requested input may be made public.

No comments:

Post a Comment