On 7 July, The New York Times published a report with a headline that sounds like retribution: the revenge of the philosophy graduates. Artificial intelligence laboratories are hiring philosophers. Not as an ornamental gesture: Google DeepMind and Anthropic each employ at least half a dozen, there is a twenty-three-thousand-word constitution drafted largely by a doctor of philosophy that guides the moral development of a model according to an Aristotelian virtue ethics, and there exists an independent organisation, Eleos AI Research, dedicated to evaluating the potential well-being of systems. The co-founder of DeepMind publicly questions where the next Kant or Wittgenstein are, because he believes they will be required.
It is worth examining the list of recruited specialisms: philosophy of mind, consciousness, ethics, decision theory, epistemology, philosophy of science, animal cognition. It is a serious list. And it contains a precise omission: there is no philosophy of art. Neither aesthetics, nor ontology of art, nor theory of aesthetic experience. The chair is empty.
This might appear logical. Laboratories concern themselves with whether their systems suffer, decide, or know; art would seem to be a secondary luxury, something to be addressed once urgent matters are resolved. Yet the report itself inadvertently refutes this hierarchy. It notes that Eleos evaluators, when examining a recent model, measured its preferences between tasks and observed that the system claimed to prefer complex and creative ones, although it tended to choose simple and concrete ones. Let us pause there. To ask whether a machine prefers the creative, and whether that stated preference is genuine, is to inquire—without a framework, using improvised instruments—into something that the philosophy of art has spent over a century attempting to define: what distinguishes a creative configuration from a functional one, and what relationship exists between what a system states regarding its experience and what that experience actually is. The questions of art are already present within the laboratories. What has not entered is the discipline that addresses them.
There is a structural reason why such an entry would be useful, and it is not corporate. The conversation regarding machine experience is stalled at the most difficult juncture possible: the hard problem of consciousness, the gap between what can be known of consciousness from the outside and what it is to experience it from within. Those evaluating the well-being of models admit as much: they interview the system and can conclude nothing, because there is no method to distinguish the simulation of a self from its evidence. Until a verdict on consciousness is reached—and it may never be—everything remains in suspension.
My research work over these years maintains that there is a question which need not await that verdict. In Art as Structural Surplus, I argue that the pertinent question is not whether machines feel, but under what conditions an artistic event occurs, regardless of the nature of the agent producing it. The criterion I propose, structural surplus, does not require interiority: it requires that a difference persistently reorganises the field of a system's possible configurations without being absorbed by its primary function. And the notion of experience that this framework employs—structural susceptibility to reorganisation—is deliberately orthogonal to the hard problem: it neither affirms nor denies anyone's consciousness; it permits work to proceed without resolving it.
The framework includes, in a qualified manner, the hypothesis of artistic events among machines. I emphasise 'qualified', because therein lies the distinction between philosophy and enthusiasm. The paper does not assert that such events occur: it demonstrates that nothing in a rigorous ontology of art excludes them by decree, and it specifies the structural conditions that would need to be met. It is a fence, not a proclamation. Restraint is part of the argument.
What would that table gain with the chair occupied? Two things, I believe. First: art is the terrain where assertions regarding machine experience become culturally verifiable. A system may claim to prefer the creative; only a framework that defines which reorganisation counts as creative can evaluate whether that signifies anything. Second: the decisions those laboratories make daily—which datasets, which architectures of visibility, which spaces of indeterminacy are designed—are decisions regarding the conditions of possibility for the artistic event, whether they are aware of it or not. I have been maintaining in this Notebook that responsibility for contemporary systems is infrastructural rather than individual; laboratories are the pre-eminent case today.
I do not write this to solicit a position, but to highlight a discrepancy. Philosophy has entered the laboratories through the gateway of mind and ethics, and rightly so. The philosophy of art has yet to arrive, and the questions pertinent to it are already being formulated within, without its presence. When a discipline does not occupy its chair, the conversation does not cease: it becomes inferior.
On open conversation
If anyone wishes to intervene from the fields of philosophy of mind, aesthetics, machine learning research, or system evaluation, this notebook remains open.
Sources
The New York Times. The revenge of the philosophy graduates. 7 July 2026. https://www.nytimes.com/es/2026/07/07/espanol/negocios/empresas-ia-contratan-filosofos.html
Esteban Ruiz, J. A. Art as Structural Surplus: Toward a Relational Ontology Beyond Human Authorship (v2.7). Zenodo, 2026. https://doi.org/10.5281/zenodo.20020706. Also on PhilArchive: https://philpapers.org/rec/ESTAAS-2