The medium is the message, McLuhan said, and the sentence has been quoted so often that it risks meaning nothing at all — which is itself worth noticing: it is exactly the kind of semantic erosion the first essay in this sequence warned against. What the phrase actually claims is narrower and more useful than its fame suggests: analysis should shift away from individual messages and toward the structures that organise communication itself, because a medium changes scale, pace and pattern regardless of what content happens to move through it. Kittler radicalises this materialism considerably by focusing on inscription technologies, on the grounds that what can be stored, transmitted and processed depends entirely on the technical systems available at a given moment; media do not simply preserve culture but actively delimit what can enter the record at all, a point that becomes crucial for artificial intelligence, where training data, file formats and model architectures together shape what can be generated and retrieved in the first place.
Bowker and Star bring classification and infrastructure together with unusual clarity, observing that categories and standards become visible above all when they fail, exclude or conflict — infrastructure, on their reading, is relationally invisible to those who fit its assumptions and highly visible to those who do not. Manovich's software studies then extend media analysis fully into computation, because cultural objects increasingly live inside software environments where interfaces, databases and algorithms determine which operations are even possible, and the old distinction between content and tool grows correspondingly unstable. Gillespie's work on algorithms and platforms closes the group by showing that ranking, moderation and recommendation all function as forms of governance: algorithms classify relevance, visibility and risk, and their power lies partly in appearing merely technical while producing a thoroughly social ordering underneath. A default is a policy because it structures behaviour at scale without ever requiring explicit coercion — a small point, but one platform critics keep having to re-argue from scratch, as if each generation of interfaces had to relearn it.
Artificial intelligence intensifies every one of these relations at once. Training data act as a kind of cultural archive; prompt histories can become genuine research traces; generative systems produce outputs whose provenance may be difficult to reconstruct after the fact; automation distributes responsibility across models, interfaces, organisations and users in ways that resist any single point of accountability. Prompts, treated seriously, are cultivated interfaces between intention and generative model, requiring iteration and documentation much like any other craft — a practice this genealogy calls PromptGarden — while the training data beneath them sit as sedimented cultural material whose historical composition keeps shaping present outputs whether or not anyone remembers where it came from. This genealogy returns, inevitably, to epistemology: search rankings influence what researchers encounter first, automated summaries can smooth over conceptual differences that actually mattered, synthetic data can produce an institutional memory detached from the events that supposedly generated it. Verification, in other words, has become inseparable from media literacy in a way earlier generations of researchers rarely had to consider. None of this licenses technological determinism — platforms never act independently of institutions, regulation or users — but knowledge now moves through systems that classify, rank, recommend, summarise and retrieve, and understanding those systems has stopped being a specialised media problem. It has become, simply, part of general epistemic competence.
Bowker, G.C. and Star, S.L. (1999) Sorting Things Out. Cambridge, MA: MIT Press.
Gillespie, T. (2018) Custodians of the Internet. New Haven: Yale University Press.
Kittler, F.A. (1999) Gramophone, Film, Typewriter. Stanford: Stanford University Press.
Manovich, L. (2001) The Language of New Media. Cambridge, MA: MIT Press.
McLuhan, M. (1964) Understanding Media. New York: McGraw-Hill.
