The Artist in the Machine: The World of AI-Powered Creativity by Arthur I. Miller

Abstract

Reviewed by: The Artist in the Machine: The World of AI-Powered Creativity by Arthur I. Miller Leah Henrickson (bio) Arthur I. Miller, The Artist in the Machine: The World of AI-Powered Creativity. Cambridge, MA: The MIT Press, 2019, 399 pp. $29.95 cloth, $19.95 paper. "Will computers ever think like us? Could they ever have flashes of inspiration like we do or come up with mad ideas? … Or do they need to? Perhaps they will function in totally other ways than human beings, come up with ideas just as great or solutions just as effective but different from the ones we would come up with" (p. xxi). Arthur I. Miller opens The Artist in the Machine: The World of AI-Powered Creativity with this provocation. Throughout the book, Miller considers what creativity actually means, and how conceptions of creative capacity may not be limited to humans. Can computers demonstrate creativity? Miller thinks so. The Artist, however, reviews various perspectives of those on the frontlines of system development. It encourages readers to consider the intersections between humans and their technologies, not just by presenting the opinion of its author, but also by welcoming readers into one of the ongoing debates about computers' contributions to culture. Miller, professor emeritus of history and philosophy of science at University College London, has been considering questions about human-technology interaction for years. In his 2014 Colliding Worlds: How Cutting-Edge Science Is Redefining Contemporary Art, for example, he considers the ways in which two seemingly disparate fields—science and art—merge for the creation of culturally meaningful aesthetic output.1 In his latest book, Miller extends his consideration of these colliding worlds to focus on one realm in particular: that of artificial intelligence (AI). The Artist in the Machine is driven by three implicit research questions: (1) What is creativity? (2) Does creativity mean the same thing for humans and computers? (3) Are computers creative? Question 1 is addressed in section I, aptly titled "Understanding Creativity." Although Miller references the work of other creativity scholars like Margaret Boden, he defines creativity as "the production of new knowledge from already existing knowledge and … accomplished by problem solving" (p. 5). Question 2 is addressed in sections II through V, wherein Miller reviews applications of AI in the realms of visual art, music, text, and musical theatre. Question 3 is addressed in section VI, "Can Computers Be Creative?" Complementing his own expertise, Miller incorporates the perspectives of AI developers throughout the book. Readers meet Google's DeepDream research team; they visit the MIT Media Lab; they pop over to Italy for a jaunt to the Fabrica research center. Miller not only discusses the functionality of AI systems; he also offers fervent insight into what goes on behind the scenes. Despite the considerable number of case study chapters (sections II through V), most of these chapters follow a rigid structure. Following a short quotation from one of its subjects, a chapter generally progresses as follows: • An introduction to a developer or development team, which focuses on providing contexts of production and personal/professional motivations for pursuing an artistic practice involving AI. • An introduction to the system created, with a short description of system functionality. • Recognition of public reception, as well as that of the developer(s), to the system under consideration. • A conclusion that shares whether or not the developer or development team believes that the system in question demonstrates creativity and whether computers can demonstrate creativity more generally. [End Page 398] This structure ensures consistent representation of systems. Although chapters are short, tending to range from five to ten pages, they efficiently and effectively offer overviews of numerous systems that exemplify varied developmental approaches. However, such a structure eventually contributes to a sense of staleness if the book is read from cover to cover. This structural aridity, though, is offset by Miller's captivating writing style. Foregoing the staidness of much academic writing, he adopts a more conversational tone. Indeed, the chapters read like pieces of narrative journalism: a scene is set, characters are developed, and Miller himself makes an occasional appearance. Technical language is explained clearly, and he skilfully abstains from digression. Further, Miller...