Some machines communicate in corporeal languages and carry meaning through sound and gesture rather than abstract associations between a word and its image.<\/p><\/blockquote><\/figure>\n<\/div>\n\n\n\n

<\/div>\n<\/div>\n\n\n\n

The material state of machines has always mimicked the sketch, and technology in general is especially poised to implement the sketch as a metaphor or method. There are obvious structures around technology-as-industry that help illustrate this: as technologies advance and become available to the average user, the old standard must be updated and adapted in order to retain functionality or, at the very least, relevance. The well-designed machine is adaptable and open-ended; much of its existence, both rhetorically and physically, is predicated on its potential to develop into something \u201cnew and improved.\u201d This is true of both hardware and software. It is interesting, then, that there is still resistance to the idea that\u200a\u2014\u200ain the same way an artist can draw a sketch, a comedy troupe can perform a sketch, or an author can write a sketch\u200a\u2014\u200aa technologist can build a sketch. Nowhere is this truer than in technological art practices, which rarely have the same incentives\u200a\u2014\u200anamely, toward capital\u200a\u2014\u200aas do the strategic plans of major tech companies. For its disruption of the unified work of art, its tendency toward and embrace of failure, and its privileging of responsiveness and adaptability, machine art is always already a sketch, rendered in aluminum and code.<\/p>\n\n\n\n

Take, for example, multidisciplinary media artist David Rokeby\u2019s The Giver of Names. <\/em>This work was first shown in the early 1990s but is still considered a work in progress, and it is updated as technology develops. (It was last shown in 2008.) In this way, it already resembles a sketch, as each iteration stages the next, with no iteration ever determined as the finished or final object. Indeed, I would argue that the concept of \u201cfinished\u201d negates the work entirely. The most recent version of the installation consists of a video camera, a computer, a monitor, a pedestal, a pile of objects, and, most importantly, an algorithm built by the artist that allows the camera and the computer to communicate with one another and the audience. To spur the algorithm to action, one or more of the objects must be placed on the pedestal, so that the camera can capture an image and feed it to the computer, thus allowing the machine to generate a sentence describing it. The language used to describe the objects is sourced from a database of words taught to the computer by Rokeby, which highlights another space in which the installation might evolve: vocabulary. When the artist decides to expose the computer to a given text in order to expand its vocabulary\/database, the question of the artist\u2019s subjectivity becomes inseparable from that of the machine\u2019s. This is a crucial disturbance of its binary logic and seemingly objective viewpoint; ultimately, The Giver of Names <\/em>can know only what Rokeby permits it to know. It is worth noting that even in the era of machine learning, robots remain susceptible to the bias of the humans that run them.<\/p>\n\n\n\n

\n

<\/div>\n\n\n\n

\n