Is AI’s Idea of Creativity Academic?
Updated: May 30
The word “academic,” with respect to the arts, has a variety of meanings and connotations today, few of them positive. Academic art is thought of as hidebound, stilted and derivative. Much of that impression is formed by the Modernist rejection of the late academic tradition (the earlier academic tradition is more complex than any critical caricature of its denouement).
An “academy” can also mean a drawn or painted study after a nude model, since the practice was fundamental to training in the figurative arts before Modernism. An artists’ academy per se, before the late eighteenth century, was a sometimes loose association of like-minded artists and architects who met to both determine official positions with respect to beauty in the arts, and to offer instruction to aspiring artists and architects. Its modus operandi qua academy was debate, discourse—Inter Silvas Academi Quarere Verum claimed Horace. The French Academy, founded in the early years of Louis XIV’s reign, dominated discourse in the arts in France until the Revolution, after which it was shut down and eventually replaced by the École des Beaux-Arts—which, it could be said, is the institution that most people today would associate with “academic” art and architecture. But, as its name reveals, the École was less of an academy than a school, albeit with a different structure than modern universities and art schools. That said, it did in fact reflect a kind of official position regarding polemical issues in the arts and architecture. And it's in architecture as the École promoted it in the second half of the nineteenth century that we find a remarkably prescient model for the idea of creativity that informs Artificial Intelligence today.
The collection of buildings, some recovered fragments and others substantial new construction, that constituted the Ècole complex (most still survive today) was a deliberate, indeed polemical, mix of historical styles and modern eclecticism.[i] What is missing in the assemblage, interestingly enough, is a representation of what might be called the pinnacle of French classicism, the architecture of the late years of Louis XIV and the reign of Louis XV, the work of Jules Hardouin-Mansart and his followers like Robert de Cotte, and Ange-Jacques Gabriel. These architects and their ateliers dominated the French Academy’s thinking in the eighteenth century, and what they were after was nothing less than a kind of perfection, an architectural ideal that could remain immutable and unchallenged. That it evaporated with the Revolution may be more a political than an aesthetic matter. But evaporate it did, and it wasn't championed at the École of the nineteenth century as the model, a French model at that, of the academic ideal.
In fact, the École wasn’t interested in perfection. At sea intellectually as a result of historicism—what the OED defines as “excessive regard for past styles”—theory was a problem of defining a modern style, not a best style as the French Academy had done. And modern, perforce, meant some sort of self-conscious extension of past styles, either a hypertrophic version of an already effusive manner as was the Second Empire’s version of French Mannerism, or some sort of eclecticism, a mélange of past styles (actually, French Mannerism was already a mélange of past and present styles). Indeed, as the century wore on, eclecticism seemed the only way forward because it was, for all its references, not a deliberate repetition of any particular past style.
Much of this methodology resulted from the aspiring architects’ extensive and comprehensive drilling in historical references, familiarity will all existing ways of organizing buildings (the parti), and thorough and meticulous documentation of historical buildings in wash rendering. This accumulation of knowledge, then, was deployed to mix, arrange, rearrange, and extend a myriad of stylistic references, from plans to façade details, toward the creation of something “new.”
Sound familiar? This is precisely what AI is doing when it takes a prompt and consults its infinite archive to assemble something “creative” from a combination of sources. It is operating precisely as an “academic” French architect would have in fin-de-siècle Paris.
In case you may have forgotten, it is exactly that facile, and often fatuous, historicist methodology that Modernist architects rejected out of hand. Their architecture, instead, would emerge from materiality, from response to program, and engage with other dimensions of modern life—especially those that were most mechanized, from steamships to automobiles.
I would argue that Renaissance and Baroque architects would have been just as critical of the École’s historicism as were the Modernists. They were interested in perfection, albeit arrived at by an adventurous rethinking of inherited models. The nineteenth-century French architects were as preoccupied with being modern as the Modernists. They just had a different way of achieving it. And of those three different ways of being creative, it is notable that software engineers have settled on the École’s eclecticism. They’ve done so because, in the end, it is the most mechanical, the most programmable, of all of them. Modernism was after something unfamiliar and defined by what it rejected, not what it imitated. Renaissance classicism was defined by what it aspired to, a tantalizing neo-Platonic ideal informed by, but not referential of, the classical past.
So, when we debate and fret over AI’s creative capacity, we would do well to recognize it for what it is, one particular way of being creative. And not, in the end, the most interesting one.
[i] The Classical and Romantic ideals which dominated French architecture during the middle third of the century met on this spot and in the process displayed their different natures. The concern with compromise that existed in the Ecole of the late 19th century was born, in part, of a controversy marked upon the building's walls. David Van Zanten, “Felix Duban and the Buildings of the Ecole des Beaux-Arts, 1832-1840,” Journal of the Society of Architectural Historians, Vol. 37, No. 3 (Oct. 1978), pp.161-174