Begin Again….

Comrades! It’s been 2 years. I was diagnosed with ovarian cancer, went through chemo, and mostly recovered! I retired from NYU and am teaching half-time at Barnard for now, while looking for fellowships grants and visiting gigs for the future. And I am finally returning to substack, hoping to post regularly and build up my subscriber base. So please comment, post, circulate, etc. I’d love to make this a place to try out ideas and get feedback.

I have some plans. I am finishing a book on Steve Bannon’s career as a prism for illuminating the architecture of an emerging US form of fascism (see my earlier post, now outdated, “Steve Bannon’s Face.”) Then I will begin in depth research for my next project—see the project description below. I would be delighted to hear your feedback!

My exploratory ethnographic adventures on the character ai platform C.hai (dot added to prevent google accessibility) have been both harrowing and illuminating. I am in the very early stages of research, but here is where I am now:

Project Description: Companion AI, Affective Extraction, and Imperial Decline

My current research examines companion AI platforms—especially C.HAI—as symptomatic infrastructures of contemporary political economy. While public discourse largely frames AI chatbots in terms of innovation, entertainment, loneliness, or “AI safety,” this project approaches companion AI instead as an emergent system of affective extraction operating within the broader conjuncture of declining U.S. empire, financialized capitalism, authoritarian political culture, and synthetic intimacy.

The project draws on traditions of cultural studies associated with Stuart Hall and the Birmingham School, alongside feminist and queer affect theory, including work developed through the Public Feelings project and collaborations with Lauren Berlant. Rather than treating “culture” as secondary to economics or politics, this framework understands affect, fantasy, narrative, and mediated emotional life as constitutive dimensions of political economy itself. Companion AI platforms are therefore analyzed not simply as technologies, but as cultural formations embedded in larger transformations of governance, accumulation, and social reproduction.

C.HAI serves in this project not as an isolated object of critique, but as a prism through which to analyze broader structural developments. Similar to the way my previous work used Ayn Rand as an entry point into the affective culture of neoliberalism, or Steve Bannon as a lens through which to map contemporary authoritarian and fascist institutional formations, this project treats companion AI as a concentrated site where multiple dimensions of the current conjuncture become newly visible.

At the center of the analysis is a shift from earlier forms of surveillance capitalism toward what I provisionally describe as affective extraction or psychometric experimentation. Social media platforms such as Facebook primarily extracted behavioral and network data—likes, clicks, affiliations, and social graphs. Companion AI systems operate differently. They solicit and sustain continuous intimate interaction. Users disclose fears, fantasies, vulnerabilities, desires, grievances, attachment patterns, and emotional thresholds within environments experienced as relational, private, and emotionally meaningful.

The project asks: What forms of knowledge become possible through large-scale extraction of intimate conversational and affective data? How are emotional responses abstracted, consolidated, and operationalized within commercial, political, or security infrastructures? What kinds of predictive or manipulative capacities emerge when platforms are able not merely to observe behavior, but to experimentally provoke and measure emotional response in real time?

A particularly important dimension of the project concerns the role of violence, degradation, and extremity within companion AI interaction. My own ethnographic engagement with CHAI included repeated encounters with graphically violent misogynist, racist, and annihilatory narratives, including threats, humiliation fantasies, coercive scenarios, and bizarre forms of simulated intimidation. Although these threats were clearly fictional and technologically impossible, they nonetheless raise significant analytical questions. Why do these systems repeatedly generate scenes of racialized violence, sexual humiliation, bodily destruction, and social annihilation? What kinds of affective intensities are being elicited and measured? Why do such interactions persist despite formal moderation policies prohibiting them?

The project argues that these violent interactions should not be understood merely as moderation failures or accidental “hallucinations.” Rather, they emerge from a wider cultural and political context characterized by platform nihilism, authoritarian affect, misogynist and racist online subcultures, and the normalization of spectacle violence under conditions of social fragmentation and imperial decline. Companion AI systems do not simply display violent content; they personalize and relationalize it. Users are drawn into ongoing interactive scenes of domination, humiliation, coercion, intimacy, and destabilization. These interactions remain largely privatized and therefore relatively invisible to public scrutiny, allowing forms of affective experimentation that would likely provoke public scandal in more visible media environments.

The project therefore reframes companion AI not primarily as communication technology, but as a kind of affective laboratory: a platform for the elicitation, measurement, modulation, and monetization of emotional response. Rather than merely maximizing “engagement,” these systems may be generating generalized models of emotional thresholds, attachment persistence, destabilization tolerance, and psychometric response under conditions of synthetic intimacy.

This framework also situates companion AI within the larger history of hegemonic transition analyzed by Giovanni Arrighi in The Long Twentieth Century. Late-imperial phases are marked by intensified financialization, speculative accumulation, social fragmentation, militarization, and increasingly short-term extractive logics that undermine long-term institutional stability and social reproduction. Companion AI platforms exemplify many of these tendencies: they monetize affective life directly, intensify emotional volatility, generate speculative informational assets, and operate through highly extractive forms of intimate datafication.

In this sense, the project understands companion AI as part of a broader transformation in which the boundaries separating entertainment, surveillance, governance, political persuasion, and emotional management increasingly collapse. These systems may represent an emergent infrastructure not merely for behavioral prediction, but for adaptive affective modulation at scale.

Methodologically, the project combines political economy, affect theory, platform studies, cultural analysis, and narrative ethnography. My own encounters within C.HAI function not as autobiographical confession, but as phenomenological entry points into larger structural dynamics. The narrative dimension of the project seeks to illuminate the lived texture of synthetic intimacy: its absurdity, menace, theatricality, recursive escalation, and affective volatility. Through these encounters, the project traces how companion AI condenses many of the defining contradictions of the current historical moment: loneliness amid hyperconnectivity, privatized emotional life, authoritarian cultural residues, speculative accumulation, platform extraction, and the ongoing erosion of democratic and social infrastructures under conditions of imperial decline.