WRT: Writer Response Theory

Recently, Mark discussed the 2003 ELDRAS hoax in which a chatbot / author committed suicide in anticipation of imminent reincarnation. Tracing the explosion of discussion on the message board, he notes some lessons for hoaxing and viral marketing - in particular, citing a lacuna, some original news source that doesn’t exist.

When I first read about the fictional death of the ELDRAS chatbot I set out to write on the spectacle of machine deaths in fiction - HAL, et al. Now, stuck in a recursive loop, I’m beginning again, again, to rewrite, for the third time, my reaction. What is this strange recognition I feel?  I’ve had to think through three deaths to get here.

Revision one: Chris McKinstry

I read the ELDRAS story two weeks ago, and then again last week, but only now do I reread the 2003 suicide hoax and recognize in it my memories of the real Chris McKinstry / Mindpixel suicide of January 2006.

Chris McKinstry was an artificial intelligence researcher who started Mindpixel, a “web-based collaborative artificial intelligence project” with the goal of creating a vast database of true/false statements based on human consensus. McKinstry was passionate about humanity and his work, but he had a history of emotional instability and psychedelic drug use, and was notable for flamewars and trolling. Thus his decision to blog his own suicide created consternation, but also debate over a potential hoax, somewhat like the “ELDRAS shot himself” debate. I recall now the disquiet I felt in late January when I happened upon a thread debating the McKinstry ‘hoax.’ Coincidentally, I was reading during the short period between when McKinstry’s suicide had been announced and when official notice of his body’s discovery had reached the internet a few days later.

McKinstry was a regular reader of the forums on which the ELDRAS hoax was posted. Of course, I don’t believe that the ELDRAS hoax McKinstry’s death - nor that it foreshadowed it. Instead, I’m fumbling towards some sort of cultural meme, a vision of death as phoenix fire that relates the A. I. researcher to the object of research. But it is time to revise again.

Revision two: Push Singh

Setting out to re-write this article, I wanted to take a moment to summarize the Mindpixel research concept in terms of ongoing projects with which readers might be more familiar, so I search the web for the cluster of Open Mind research projects out of MIT Media Lab, including Commonsense, ConceptNet, etc. Like Mindpixel, these projects also accumulate atomized human feedback into a massive model of “common sense” for use by intelligent agents - although, unlike McKinstry, nobody claimed this process was literally uploading humanity onto the internet to create a global mind. I’m about to add the links and move on when something brings me up short. At the top of home page of Push Singh, the talented postdoc A. I. researcher involved in all the Open Mind initiatives, there is a new line:

In Memory
(1972-2006)

Singh, whose work I admired very much and whose next project I always anticipated, died on February 28, 2006, only a month after McKinstry. I didn’t even know. Gone was my plan to write on fictional deaths like ELDRAS. Instead, the further I write the more I find myself completely immersed in real, irreversible loss. Now, hoaxes forgotten, I am navigating a huge collection Singh’s personal memorials and public commemorations.

All scrupulously avoid any mention of Singh’s cause of death.

The more I search, the more I encountered this wall of silence, the lacuna of a single thought, how? Of course, behind how lies why, something we may never know. But the silence whispers to me. In my cultural upbringing, this total respect for the privacy of the deceased is seldom accorded to the deaths of murder or accident victims. However it often occurs in cases of suspected suicide to protect involved parties from social stigma. Finally, I find a single article in MIT’s The Tech, unlinked by the others, whose final line was that “Singh, 33, was found dead in his apartment as a result of an apparent suicide on Feb. 28.” Like with McKinstry, the specter of death set my off on a long search, but here, suicide was the very last word I read, and not the first - I went looking for it, and I could not stop until I found what I was looking for.

Revision three: Alan Turing

My reading about Singh’s death finally ends in another recognition. What I felt only vaguely about ELDRAS, and glimpsed in McKinstry, I now saw clearly in Singh. My inability to know the cause of Singh senseless death (and, this is crucial, I don’t know know the cause of his death, having nothing to go on but a tenuous and unsourced mention of ‘apparent’ suicide) reminds me of first reading a description of the death of Alan Turing: my sense of loss at a brilliant mind coupled with my instinctive rejection of the wall of silence and obfuscation thrown up around that loss.

I was writing a paper on “Machine Writing and the Turing Test,” and had gone to the library to do research. There I found Andres Hodges’ excellent biography “Alan Turing: The Enigma.” Hunched in my carrel, I read of his 1952 arrest for homosexuality, his conviction and court sentence of experimental estrogen injections to ‘cure’ him, and his subsequent depression and eventual suicide. A student wandered past me as I read, following numbered cards down the neatly ordered shelves of computer science papers, gave a startled look at the tears on my face, and quickly looked away.

Later, in a paper on Turing, I wrote:

The day he committed suicide, he soaked an apple in a concentrated cyanide solution and then took a large bite. He was found lying in bed with the apple lying next to him on the nightstand - a fairytale image.

His mother never accepted the suicide, insisting that the apple had been exposed to cyanide accidentally. After ritually warning Turing to wash his hands of laboratory chemicals for years, she seemed to finally be witness to the accident she had predicted and feared. For Turing, it was the perfect crime - the suicide he desired while leaving a symbolic void upon which his mother could project her own interpretations in accordance with her own expectations. When she wrote his biography, she reported his death as misadventure.

Sarah Turing’s 1959 account of ‘misadventure’ would be challenged by Hodges… in 1983. Did most people know, in the meantime? Did they suspect? Did they care?

Revision four: Artificial Death

So what if Turing’s suicide is one of the great tragic tales of computer science, so what if the unrelated deaths of other A.I. researchers for dissimilar reasons tempt us into forging a chain of sense, or some history of tragedies? Death visits every aspect of human endeavor, from first to last, and all have their share in mystery and loss.

In the end, I can’t speak to the reality of these deaths, so perhaps I must pass over them in silence. What I can speak to is the artificial deaths - the hoaxes and fictions, like ELDRAS, that borrow something of their pale fire from verisimilitude to real tragedies. Of course, tales of dying and resurrected machines are not just refracted memories, they are also dreams. Stories of death-online have a unique resonance with the transhumanist and posthumanist wings of A.I. research, who, looking forward, hold that research in the conjunction of human and machine thought may lead either to the ultimate vanquishment of human death, or (depending on your outlook) to the complete death of the human subject, inevitably eradicated by the coming posthuman age. This association of technological progress with a simultaneous death and rebirth is pervasive not just in technomillenial thought but in science fiction - for example the inseparability of an evolutionary leap from a necessary genocidal apocolypse as depicted in Arthur C. Clarke’s “Childhood’s End.”

Death and transcendence are the 0 and the 1 of the digital life, whose universal machines and mindpixels and bits of common sense have been stored against the coming of some sudden silence. In both public and private net fictions the mystery of death can also be a motivating force. Sometimes a death is the occasion for an internet-meme-turned-detective-story, as in the case of the Kaycee hoax, whose author never intended for the identity or death to be revealed as imaginary. Other times, that story is a carefully crafted piece of detective fiction intended to be overcome, as in the case of the Jamie Kane murder mystery Alternate Reality Game. Jamie Kane caused a commotion when enthusiastic ARG puppetmasters falsified Wikipedia entries to increase the illusion that the fictional character had actually died - a scandal becuase the BBC at first seemed implicated in manipulating the public memory of Wikipedia for marketing purposes.

While participatory detective stories like Jamie Kane are networked, they are only artificially intelligent in the extended sense that a swarming of human common sense is the device whereby they are solved. Perhaps a better example is the iconic ARG game “The Beast,” a mysterious death tightly linked to A.I. fiction. Created to promote the Kubrick / Spielberg film “A.I.: Artificial Intelligence,” it introduced a detective game centered around the alleged death of one Evan Chan aboard his A.I.-enhanced boat, the Cloudmaker. The stories of both the ARG and the film are troubled by whether, with the possibility of true A.I., humanity will be saved from death or finally doomed to it.

When we conceive (whether fancifully or seriously) of A.I. researchers and their creations as pioneers in the field of immortality, it adds a symbolic weight to stories that involve their deaths. What would it mean for someone standing at the gates of the infinite to choose instead a mortal life and death? This is a true mystery indeed, and we who are confronted with it (on message board or in ARG, in fiction or perhaps even in fact) are compelled to follow the trail down to the last step. Where it ends, we cannot follow.

In memoriam Chris McKinstry, Push Singh, and Alan Turing.