This book got mixed reviews (as summarised in the first paragraph of my review) but I found it a thoroughly enjoyable entry into the world of robotics and real-life androids. I was considering posting this on my other blog but thought that the themes of personhood and authenticity are perhaps more germane here than I thought. Indeed, I am gearing up for posting something more explicitly philosophical in tone than this.
I also had read only some of Iain McGilchrist’s “The Master and His Emissary” at the time I wrote this review, which didn’t stop me referencing it (and emailing McGilchrist to ask him what he thought of the Turing Test) More recently I read the whole thing and … I was right. So there you go. I’m afraid I may have perpetrated this kind of thing in the past a little bit.
After all, I have read How To Talk About Books You Haven’t Read . I really have.
Anywhere, here is the review:
How to Build an Android:
The True Story of Philip K. Dick’s Robotic Resurrection
David F. Dufty
Henry Holt and Company, 273 pages
Just over seven years ago, the head of Philip K Dick went missing from an America West Airlines flight between Dallas and Las Vegas. A tired roboticist, transferring the talking robotic replication of Dick’s head from one tech presentation to another, left it in an overhead baggage locker. An incident which has already inspired a radio play (Gregory Whitehead’s Bring Me The Head of Philip K Dick) and received substantial media coverage at the time, it initially seemed to me somewhat too slight to merit book-length treatment. Perhaps a long piece in Wired would do it justice. And indeed, surveying what other reviewers have made of the book (David F. Dufty has handily compiled prior reviews, including poor ones, on his website), I find that some have concluded with my initial impression. For instance, New Scientist‘s Sally Adee found “50 pages of detailed historical introductions to every last person involved in the android project… Dufty recounts conversations in exhausting detail, and finds nothing too small or insignificant to share with the reader: we learn where the Starbucks is at several convention centres, we learn of one room that “the frame was made out of timber.” We learn that Google created a famous search engine.”
I however found Adee’s criticism unfair, and somewhat beside the point. Dufty, a postdoc in the University of Memphis at the same time as many of the events described and therefore working with many of the personalities involved, has crafted a readable narrative which ranges from the nature of academic politics (and the grant applications that take up most of any senior researchers time) to the distinctions between Alan Turing’s and Philip K Dick’s visions of what distinguishes — or could distinguish — computers from humans. In the end, the book dealt with weighty themes, some of the weightiest themes we can think of. As Henry Markham of the Blue Brain project so eloquently describes in his TED talk on the subject, computational simulation of the human brain is one of the grandest challenges we can conceive (and possibly an unattainable one, although that’s another debate) Dufty may have a somewhat flat, deadpan style, but it reminded me of the dictum (possibly one of Robert Louis Stevenson’s) that extraordinary narratives should have an unadorned, simple style.
If the book has a protagonist, it is the man who left the head in the overhead luggage compartment on that fateful flight, David Hanson, a trained sculptor turned roboticist who passionately argued — contra to the prevailing wisdom in the robotics community that aesthetics don’t matter — that beautiful and lifelike humanoid robots were crucial in the development of robots that would truly revolutionise our lives. Hanson emphatically rejected the notion of the “uncanny valley,” the supposed phenomenon whereby, as robotic models and digital representations of humans come closer and closer to being lifelike (while missing the mark slightly), we are more and more repulsed. Intuitively the uncanny valley makes sense to many, yet as Hanson has pointed out there is a lack of empirical evidence to support its existence.
Artificial intelligence has evolved to become focused on specific tasks, often those of intellectual prowess (such as beating Garry Kasparov at chess) rather than the overall simulation of human mental functioning in all its manifestations. This has lead to great, headline-catching successes (such as beating Garry Kasparov at chess) but has arguably lead away from a visionary, transformational view of the possibilities of AI. Hanson advocates approach to robotics grounded more in a gestalt view of humanity and human-ness than the mere performance of tasks in isolation, and one which emphasises the aesthetic nature of the whole android concept. For Hanson, leaps of scientific progress are as much artistic and aesthetic as anything else. Dufty describes the combination of sculpting craft and high tech that goes into the creation of a Hanson style robot very well.
Philip K Dick was an ideal candidate for potential immortalisation as a robot head in many ways. Obviously, his fiction had dealt explicitly with themes of humanity and humanoid robots, and the difficulty distinguishing between them. Empathy, rather than Turing’s imitable intellectual functioning, was the key. Dick has become more than a cult figure and is now widely regarded as a key American author of the second half of the Twentieth Century. Any Dick related project will garner attention, and the project coincided with the production of the Richard Linklater film A Scanner Darkly, and indeed was identified as a publicity aid for the film. Also, Dick’s reputation as a sort of neo-gnostic eccentric meant that elliptical or cryptic responses which might otherwise be seen as failures of artificial intelligence would be seen as just typical Philip K Dick.
Another characteristic of Dick made him an ideal subject for such a project. Although he was dead and therefore his head couldn’t be directly modelled from life, there was a vast archive of conversations he had had with all comers in his California bungalow in the 70s, when his house had been a sort of perpetual symposium of dropouts and outcasts with whom he would hold court. These conversations covered a vast range of topics, esoteric and everyday, which allowed the team to create a bank of possible responses to a great deal of questions. They also programmed some standard responses to questions such as “what is your name?” They never programmed Dick with a response to “do androids dream of electric sheep?”
The head was a hit at the various conferences and exhibitions it was displayed at, to the extent that each member of the public who patiently queued up to meet it could only have a minute or two of interaction. Dick’s daughters were consulted about the project, and after being convinced of the good intentions of those involved agree, but had an understandable ambivalence about it. The head did tend to get caught in infinite verbal loops, which the roboticists tried to manage by creating a “kill switch” to terminate logorrheic conversations. In its exhibited life the head was, to a certain degree, something of a Mechanical Turk, with a human behind the scenes desperately trying to maintain the illusion of spontaneous conversation.
I was reading the English psychiatrist Iain McGilchrist’s The Master and His Emissary around the same time as Dufty’s book. McGilchrist’s book is a massive, sweeping, visionary book which argues that the division between the two hemispheres of the brain — the one which is grossly simplified into the dichotomy of logical left brain and creative right brain — has been not only a determinant of human history and culture but THE great determinant. McGilchrist has marshalled an enormously impressive range of philosophical, empirical, artistic and other forms of evidence for his argument, and while it is not utterly persuasive in all respects and hemispherical specialisation is itself far from a binary, dichotomous phenomenon, it is a book worth arguing with. In any case, McGilchrist again and again assails what he terms the left-brain tendency towards decontextualized analysis and away from an appreciation of holistic and of nuance. Artificial Intelligence’s turn to a task-focused approach is, in McGilchrist’s terms, a classic case of the triumph of the left brain.
Dufty’s book is deceptive. Initially it seems a rather bald account of the story of Dick’s head, but it builds into a thought-provoking book. Dufty marries the exciting, speculative world of contemporary AI and robotics with the prosaic reality of grant applications and presentations at noisy, busy, conferences. There is an amusing thread of Talking Heads references throughout — indeed David Byrne is a not insignificant player in the story . One of these references is slightly off the mark though — while Talking Heads did do a version of “Take Me To The River,” it is originally an Al Green song. Why does this come up? You’ll have to read the book to see.