We're at a cultural moment in which machine-generated work doesn't merely resemble human creation but actively rewrites the criteria by which authenticity itself is assessed.

A new paper in AI & Society (Springer, 2025) makes this argument, and proposes that prevailing frameworks for understanding artificial intelligence (as a tool, a medium, or an accelerant of existing cultural processes) are theoretically insufficient.

"Synthetic Media and Computational Capitalism: Towards a Critical Theory of Artificial Intelligence," develops three interlocking concepts to describe what it calls the algorithmic condition: a cultural moment in which machine-generated work does not merely resemble human creation but actively rewrites the criteria by which authenticity itself is assessed.

The Inversion

The paper's central concept is the Inversion.

Where earlier critical analyses of AI focused on automation (machines replicating existing cultural labour), the Inversion describes something more disruptive: a reversal in which synthetic production no longer just simulates human output but begins to set the standard against which human output is measured.

The result is that "reality itself," in the paper's terms, is no longer authenticated through traditional social mechanisms (e.g. institutional trust, authorial identity, shared verification) but through computational processes that operate prior to and independently of those mechanisms.

Automimetric production

The second contribution is the concept of automimetric production: a framework for understanding how AI systems create value not by copying pre-existing forms but by generating new cultural objects that inherit the appearance of human authorship without its conditions.

This is distinguished from prior theories of simulation (Baudrillard is an implicit interlocutor) by its emphasis on political economy.

Automimetric production is situated within computational capitalism, a social formation in which algorithmic systems restructure access to authentic experience and the capacity to verify claims about the world.

From false consciousness to post-consciousness

The paper extends the Frankfurt School tradition by arguing that the classic concept of false consciousness, the ideology that obscures real conditions from subjects who nonetheless remain subjects, is inadequate to the algorithmic condition.

In its place, the paper proposes post-consciousness: a state in which the boundary between individual and synthetic consciousness becomes genuinely porous, not merely ideologically obscured.

This is not a claim about subjective experience being "taken over" by machines, but a structural argument about how the production of meaning has shifted such that the distinction between human and algorithmic genesis can no longer be reliably maintained from within ordinary cultural experience.

Constellational analysis and algorithmic détournement

Methodologically, the paper proposes constellational analysis, drawn from Adorno and Benjamin, as a way of mapping the interplay between technical systems, cultural forms, and political-economic structures without reducing one to the others.

It concludes with a call for algorithmic détournement: practices that redirect computational systems toward human flourishing rather than instrumental rationality, borrowing the Situationist concept of repurposing dominant cultural forms against their own logic.

Why this matters

The paper's challenge to celebratory "AI as new medium" framings is the most theoretically developed of its kind in recent critical literature on the subject.

Where optimistic accounts treat generative AI as an extension of prior media revolutions (photography, cinema, the internet) this framework insists that the Inversion represents a categorical break: not a new tool for human expression but a restructuring of the conditions under which expression, authorship, and shared reality are constituted at all.

The verification crisis this produces, the paper argues, compounds existing strains on institutional authority and social trust, and risks new forms of stratification based on differential access to human experience and the capacity to authenticate it.

The paper is open access and available via Springer's AI & Society journal.