AI, ChatGPT, Life, Environment, and Empire

Abstract:

This talk emerged from the convergence of three situations in one eventful week in October.

First, my MSc students rebelled when I tried to advise them to use Generative AI to learn how to code. Second, I did a careful read of Karen Hao's 2025 landmark "Empire of AI" book. Third, I attended a workshop at the Exeter Critical AI Research Centre together with several of the leading British scholars engaged in radical critique of AI.

I had not realised until the convergence of these three situations that the framing of critical AI discourse is in urgent need of a shift.

In recent years and for decades prior, critical AI scholars (myself included) have formulated deflationary definitions and arguments about AI. The general narrative amongst critical scholars has been that AI is nothing much special. In past years AI was an unattainable, shifting, always future-looking imaginary. Now Generative AI is frequently described as "just statistical text prediction".

The problem with these types of narrative is that there is then no reason to be especially concerned about AI in itself. Such critiques of AI rely on general anti-capitalist, anti-imperialist, or neo-luddite lines.

Do any specificities of AI necessitate unique lines of critique?

In this talk I am to offer both a new strategy for deflating AI hype and a new strategy for leaning into it. I present both as refinements of Hao's arguments.

I propose to define AI simply as ChatGPT and its AI assistant platform ilk.

I then suppose that some of the tallest claims about AI are true---that AI presents us the "third industrial revolution"---and examine implications of that.

Peaks Krafft, Lecturer in Computational Sociology

Chair: Shaira Vadasaria, Senior Lecturer in Race and Decolonial Studies