By Oskar Galeev (Johns Hopkins University)

Artificial General Intelligence is increasingly treated as a unified imaginary, a technology to which contemporary discourse routinely attributes the predicates of omniscience (it knows all data), omnipresence (it is everywhere via cyberspace), and omnipotence (it could save or destroy everything). This paper argues that AGI, so understood, is not yet a technological artifact but a referent that fuses three conceptual pillars into a single framework: timelines, which recast the future as a measurable countdown; alignment, which approaches the future as an engineering problem of control; and existential risk, which casts the future as a horizon of irreversible outcomes. Each pillar carries its own temporality, and their simultaneous hold on imagination is what gives AGI discourse its traction.

The pillars formed within an intellectual lineage that philosophy of technology has mostly read as ideology: the trajectory from 1970s–80s cyberutopianism through the Extropian and transhumanist fora of the 1990s to the consolidation of rationalism and Effective Altruism schools in the 2000s–2010s. It was within these traditions that the vocabulary now organizing AGI debate was produced: intelligence explosion, orthogonality, instrumental convergence, superintelligence. The paper reads AGI discourse as a fusion of techno-philosophical and techno-theological thought — a system of beliefs that imbues technological systems with transcendent significance, and one that requires reconsidering how philosophy of technology presumes the relationship between belief and techne. A critical reading of how this lineage became the dominant framework for thinking AGI also reveals what it has excluded: alternative framings of AI ethics, in particular, have been displaced by the existential-risk register.

The paper draws on fieldwork in AI governance and existential risk organizations, interviews with key figures in the Extropian-to-EA milieu, and archival work on the movement's formative materials, all forming part of a dissertation on the intellectual history of AGI. It aims to demonstrate not only that a critical philosophy of human–technology relations needs to reckon with the transhumanist philosophies that shaped AGI discourse, but also that this reckoning is itself a site of inherent tension within philosophy of technology.

By Diana Lengua (University of Essex)

Much contemporary philosophy of AI remains structured by questions inherited from Western metaphysics: Is AI an object, a tool, an agent, a partner, or an infrastructure? Does it think, act, decide, or just compute? These questions matter, but they often presuppose a model of technology as something to be classified according to its proximity to human agency.

I argue for a comparative theory of technogenesis to move past models that start with isolated entities and then ask how technology extends, replaces or threatens them. This theoretical approach allows AI to be read as a philosophical event not because it produces a new type of mind, but because it exposes the limits of the inherited categories through which mind, agency, technics, and world have been separated.

The paper contributes to debates in the philosophy of technology by asking whether AI can provincialise the field itself. Its central claim is that AI should be understood as a site of comparative technogenesis where traditions of technology and ontologies of relation are placed under productive tension.

By Kenneth Howarth (Mercer County Community College)

Artificial Intelligence is commonly described as a tool, agent, infrastructure, partner, object, or socio-technical assemblage. Each description captures something important, but each becomes misleading when treated as exhaustive. This paper argues that Artificial Intelligence should be understood, for many philosophical and practical purposes, as a judgment-shaping technic: a technological configuration that redistributes salience, attention, sufficiency, responsibility, and contestability within human and institutional life.

The argument responds to a central question in contemporary philosophy of technology: how should inherited categories of technics and human–technology relations be reconsidered in light of Artificial Intelligence? Rather than selecting one status for Artificial Intelligence in advance, I propose a bearings-oriented approach. By “bearings,” I mean the load-carrying orientations through which agents and institutions discern what matters, what counts as enough, and what warrants action practically and ethically. By “moorings,” I mean the anchoring, ballasting, tethering, and revising practices that keep action from becoming ungrounded, overconfident, or merely accelerated.

Artificial Intelligence affects both bearings and moorings. It can support inquiry, extend perception, and improve coordination; it can also compress deliberation, over-stabilize appearances of adequacy, and shift responsibility away from the sites where judgment is actually being formed. The question, then, is not only whether Artificial Intelligence acts, understands, or possesses agency. The more immediate philosophical question is how Artificial Intelligence reorganizes the conditions under which humans and institutions experience action as warranted.

This approach draws on pragmatist, phenomenological, and postphenomenological philosophy of technology, while also engaging recent debates about machine ontology and moral status. It accepts that Artificial Intelligence can appear differently as object, tool, milieu, agent, infrastructure, or partner depending on the operative setting. But it resists both model-file reduction, which treats Artificial Intelligence as a bounded technical artifact, and assemblage dissolution, which expands Artificial Intelligence so widely that responsibility becomes difficult to locate.

The paper concludes that Artificial Intelligence is not one thing for all purposes, but neither is it whatever discourse makes of it. It is a situated technic whose individuation should preserve explanation, responsibility, contestability, and revision under uncertainty. A bearings-oriented account therefore reframes Artificial Intelligence not as a self-authorizing replacement for human judgment, but as a powerful reconfiguration of the relations through which judgment is formed, delegated, challenged, and sustained.