Environmental Intelligence: Context and Rhetoric

Technologies appear unquestionably thing-like and solid.They seem so obviously part of the furniture of the world.We tend to unthinkingly adjust our ontological outlook to grant them such a solid existence.But we need not.Sabina Leonelli's (2025) admirable paper continues a line of critique and reconceptualization of the newish technology of AI in challenging such lazy philosophical conclusions.A third of a century ago Glamour (1988, p. 195) already observed "Artificial intelligence is philosophical explication turned into computer programs."Nowadays, we might say 'AI is philosophy (epistemology, ontology, and ethics) turned into business models via the mediation of computer programs and appropriated data.' Yes, as she exhorts, philosophy does matter.Leonelli's reframing of AI as environmental intelligence (EI) aligns well with earlier reconceptualizations of technologies.She focuses upon the "embodied and situated nature of intelligence."As she observes (in her footnote 2), her critique depends on the notion of "environment," which is laden with many meanings.To sidestep that baggage, I will focus upon environment as context, 1 by which I mean all the stuff that is other.Thus, my environment contains everything that is not me.Your environment, in contrast, includes me as part of it.The environmental perspective on AI means we pay attention to everything outside of the highly guarded black box itself, which most importantly includes the physical resources to make the technology work (as Leonelli describes), but also the data on which AI is built, and the rest of the world, including people, on which it acts and affects.Earlier, the nascent field of cybernetics (from which AI grew) was reconceptualized environmentally as 'second-order cybernetics' where the cybernetic system-the original black box (von Hilgers, 2011)-was viewed embedded into a richer societal context (Scott, 2004).There is now a growing literature on how "AI fails us" (Siddarth et al., 2021), which notes it is usually framed as "a wasteful imitation of human comparative advantages and a confused vision of autonomous intelligence, leading it toward inefficient and harmful centralized architectures" (p. 1).They opine "we are not locked into this trajectory" (p. 1).Indeed, we are not; as Mumford (1944) observed 80 years ago, "Nothing that man has created is outside his capacity to change, to remold, to supplant, or to destroy: his machines are no more sacred or substantial that the dreams in which they originated."An 'environmental' perspective is now common in scholarship on other technologies.After all, one can hardly understand automobiles (for example) by just looking at the artefact (the car itself) without also paying attention to the roads (who pays for them, what buildings are knocked down in the process), and all the other infrastructure necessary for their widespread use (fuel acquisition, refining, distribution, and sale; maintenance and repair of cars; hospitals, for the inevitable accidents); the pollution caused; the traffic jams; the positionality of goods (cars as status symbols, not as mere transport), and so on-in other words, the whole sociotechnical system.