Miller, Arthur I. The artist in the machine: the world of AI-powered creativity. Cambridge, MA: MIT Press, 2019. xviii, 399 p. ISBN 978-0-262-04285-7. $29.95.
Before I started writing the review, I visited the Website devoted to this book www.artistinthemachine.net. My computer could not open the 'audios and videos' part for some reason, but it is not a big problem as you can trace most of the projects described in the book on the Web anyway. So, I have been lost in following them up and being frustrated when some were not doing what I was led to expect, e.g. the
The fact that the author has collected all this interesting work of art and interviewed their creators is remarkable and shows his devotion to the topic. He has also presented his material in a very accessible and readable text. The chapters from II to V can be read as a wonderful science fiction story with the only difference that it is not fiction at all.
The main subject matter of the book (at least for me) is the artistic work of modern artists who use computers and artificial intelligence as the tools in very creative ways. These artists combine higher or, very seldom, lower competence of programming and computing with more traditional instruments of trade, such as words, colours and sounds or the mix of them. Some of them regard their computers or software as tools, but others as collaborators and co-authors, a minority places the whole creative process and authorship with the technology. Sometimes, the border between the artist and the machine is really difficult to distinguish, though I do not share the opinion that the computers are creating art all by themselves. As the author points out at the end there are very many necessary conditions missing, at least for now. Nevertheless, the potential of new technologies for creative artistic work is demonstrated very persuasively in this book.
The book is divided in six parts, the first of which is discussing the concept of creativity and the final one reflects on creativity of computers. The four middle parts present the actual study done by Miller. It involves different cases of creative projects and interviews with their artists.
The second part is devoted to visual arts and different projects from the 'simplest' mimicking painting styles or techniques, to those, where computers produce entirely new art objects. The third follows the musical compositions produced by computers with very interesting ventures into robot jazz band, folk music, popular songs or dodecaphonic pieces. The fourth presents the projects related to 'magic with words': mainly poetry, but also humour and even a screen script. The fifth is just about one, incredible project of creating and staging a computer-aided musical 'Beyond the Fence'. (Here you can find the fragments of it on the YouTube). The whole process of creation and staging unfolds for the reader. I will leave for you to judge what it says about the role of computers and humans and whether you would see it as a cooperation between the actants.
I was not sure what or who is the emphasis of the study. It seemed all along that the emphasis is on the computers and their creativity. However, the author painstakingly presents the biographies of each artist, writes very detailed account of their working techniques and respectfully relates their opinions of art and computers. If you add that the author on several occasions talks about the role of human interpreters reading the computer poetry, acting in the computer-scripted film, or selecting the options produced by computers, it will be very difficult to see any software or technology as anything else but a very advanced tool. However, this is the whole point of the book and its last chapter - engaging in an interesting intellectual discussion and driving ahead the development of information technology.
I appreciated the book very much, but there are some features of it that make it less attractive. The first chapter discusses several approaches to creativity. However, the author suggest a quite clear picture of seven hallmarks, two marks of genius, three types and four stages of creativity. Some of them are based on the author's studies and some on other researcher's results. In any case, all this counting seems suspicious and leads away from a fruitful discussion of creativity as a very complex element of human activity. Some lists are not entirely well argued. For example, the seven hallmarks may have nothing to do with creativity, but simply signify success in certain activities, which is not the same at all. Having clearly cut, though quite complex, lists helps in arguing the main thesis about computer creativity, but does it help understanding it as a way to what is 'new, surprising and valuable' (p. 25)?
The author also provides quite many examples of creative people, who all happen to be successful white men. It seems that women can only be creative if they are spouses (Marie Curie and Sylvia Plath) or lovers (Anaïs Nin) collaborating with creative men (p. 12). There are no examples of any creative people of colour, children, or anyone else. The following parts provide a more diverse picture of creativity among artists, so it is not quite clear why this diversity is forgotten in the first part. I guess this omission was not intentional, but it is alarming nevertheless. Another alarming issue, was the number of creative people working for big technology companies. It is clear that driving some of these very expensive and highly technological projects requires resources that only the likes of Google or Amazon can provide. But I do not believe that they support this development because of the love of theatre or poetry.
The author presents the development of creativity of computers through the achievements of programming in playing sophisticated games. Machine learning, neural networks and information processing have extremely advanced artificial intelligence and computer powers. There is no question that even more powerful methods and processes will emerge in the future. They definitely have potential in modelling the activities of human brain and understand some of its work. But as we do not really know for sure what, how and to what extent influences human brain, it is difficult to say if this technology relates to anything close to human creativity. We have to remember that computers are not social and natural beings. We know for ages that our minds are the products of our societies, not only natural evolution. On the other hand, we just now start discovering the impact of our gut bacteria on mind work (see, e.g., Sherwin et al., 2019). It is difficult to imagine computers developing cravings for pickled gerkhins or suffering from indigestion even if we have depicted technological and robot societies. There can be many arguments for and against creativity of computers and people. Therefore, this book and its final part are very useful in encouraging one to think about these interesting and important questions.
The readers can understand that I was not convinced by the author of computer artistic creativity, but I have much appreciated his very thoughtful and intelligent approach to the artwork created through them and the artists that he obviously sees as the driving creative force behind technology. The book would be interesting for many different groups of readers in technology and humanities, philosophy and information science, arts and neuroscience. It is worth reading and thinking about the issues raised in it.
Reference
Sherwin, E., Bordenstein, S.R., Quinn, J.L. Dinan, T.G. & Cryan, J.F.. (2019). Microbiota and social brain. Science, 366(6465), eaar2016. DOI: 10.1126/science.aar2016
Elena Maceviciute
University of Borås
May, 2020
How to cite this review
Maceviciute, E. (2020). Review of: Miller, Arthur I. The artist in the machine: The world of AI-powered creativity. The MIT Press, 2019. Information Research, 25(2), review no. R688 [Retrieved from http://www.informationr.net/ir/reviews/revs688.html]
Information Research is published four times a year by the University of Borås, Allégatan 1, 501 90 Borås, Sweden.