A world model: On the political logics of generative AI

From OpenAlex, which stores abstracts as an index of words and their positions: what is shown is rebuilt from that, and reads fluently whether or not it is what the author wrote.

Abstract

The computational logics of large language models (LLMs) or generative AIfrom the early models of CLIP and BERT to the explosion of text and image generation via ChatGPT and DALL-Eare increasingly penetrating the social and political world.Not merely in the direct sense that generative AI models are being deployed to govern difficult problems, whether decisions on the battlefield or responses to pandemic, but also because generative AI is shaping and delimiting the political parameters of what can be known and actioned in the world.Contra the promise of a generalizable "world model" in computer science, the article addresses how and why generative AI gives rise to a model of the world, and with it a set of political logics and governing rationalities that have profound and enduring effects on how we live today.The article traces the genealogies of generative AI models, how they have come into being, and why some concepts and techniques that animate these models become durable forms of knowledge that actively shape the world, even long after a specific material commercial GPT model has moved on to a new iteration.Though generative AI retains significant traces of former scientific and computational regimesin statistical practices, probabilistic knowledge, and so onit is also dislocating epistemological arrangements and opening them to novel ways of perceiving, characterising, classifying, and knowing the world.Four defining aspects of the political logic of generative AI are elaborated: i) generativity as something more than the capacity to generate image or text outputs, so that a generative logic acts upon the world understood as estimates of "underlying distributions" in data; ii) latency as a political logic of compression in which (by contrast with claims to reduction or distortion) the thing that is hidden, unknown or latent becomes surfaced and amenable to being governed; iii) broken and parallelized sequences as the ordering device of the political logic of generative AI, where attention frameworks radically change the possibilities for governing non-linear problems; iv) pretraining and fine-tuning as a computational logic of generative AI that simultaneously shapes a "zero shot politics" oriented towards unencountered data and new tasks.Across each of the four aspects, the article maps the emerging contemporary political logic of generative AI.With a single, configurable world model engine, rather than a separate model for every situation, knowledge about how the world works may be shared across tasks.(LeCun, 2022: 5).Knowledge takes up residence in a new space […] What event, what law do they obey, these mutations that suddenly decide that things are no longer perceived, described, expressed, characterised, classified, and known in the same way?(Foucault, 2003: 235-6).In a 2022 essay, the Turing laureate computer scientist, Yann LeCun, set out a "path towards intelligent machines" in which he draws a familiar distinction between how humans and machines respond to what he calls an "unencountered situation" (2022: 1).Humans and animals, LeCun proposed, "know how to act in many situations they have never encountered".For example, they do not need to encounter all possible situations where something may be hot, or where an object might fall, in order to know and to act in advance of other future potential hot or falling objects.In machine learning, by contrast, "even the rarest combination of situations" must be "encountered frequently in training"every potentially hot or falling object has to be included in a training datasetif the system is to "avoid making dangerous mistakes when facing an unknown situation" (LeCun, 2022, p. 3).For LeCun, as for many other commentators, the pathway of future AI is defined by a new paradigm of knowledge under conditions of uncertainty, one where a machine learning model must be able to discover something in a data distribution, something essentially generalizable to new and unseen situations.This orientation of AI towards general discovery of "how the world works" and adaptation to new domains and tasks is captured by what LeCun calls a "world model", which would supply "an internal model of how the world works" so that AI becomes "configurable" to each new situation it encounters (2022: 2-3).The powerful claim that a flexible, reconfigurable world model could deal with all potential future unencountered situations defines much of the politics of contemporary generative AI.It is a claim that promises a general resolution of difficult problems across technical computational and political paradigms: an AI model that draws upon a structure of "how the world works" in order to respond to an input it had never encountered in training; and a political model that is always capable of action in the face of the unencountered