Living with the Genie: Essays on Technology and the Quest for Human Masteryby Alan Lightman (Editor), Daniel Sarewitz (Editor), Christina Desser (Editor)
<p>Biotechnology…Cloning…Robotics…Nanotechnology…..<p>At a time when scientific and technological breakthroughs keep our eyes focused on the latest software upgrades or the newest cell-phone wizardry, a group of today’s most innovative thinkers are looking beyond the horizon to explore both the promise and the peril of our technological future.<p>Human ingenuity has granted us a world of unprecedented personal power—enabling us to communicate instantaneously with anyone anywhere on the globe, to transport ourselves in both real and virtual worlds to distant places with ease, to fill our bellies with engineered commodities once available to only a privileged elite.<p>Through our technologies, we have sought to free ourselves from the shackles of nature and become its master. Yet science and technology continually transform our experience and society in ways that often seem to be beyond our control. Today, different areas of research and innovation are advancing synergistically, multiplying the rate and magnitude of technological and societal change, with consequences that no one can predict. Living with the Genie explores the origins, nature, and meaning of such change, and our capacity to govern it. As the power of technology continues to accelerate, who, this book asks, will be the master of whom?<p>In Living with the Genie, leading writers and thinkers come together to confront this question from many perspectives, including:<ul><li> Richard Powers’ whimsical investigation of the limits of artificial intelligence<li> Philip Kitcher’s confrontation of the moral implications of science<li> Richard Rhodes’ exploration of the role of technology in reducing violence<li> Shiv Visvanathan’s analysis of technology’s genocidal potential<li> Lori Andrews’ insights into the quest for human genetic enhancement<li> Alan Lightman’s reflections on how technology changes the experience of our humanness<p>These and ten other provocative essays open the door to a new dialogue on how, in the quest for human mastery, technology may be changing what it means to be human, in ways we scarcely comprehend.
- Island Press
- Publication date:
- Edition description:
- Product dimensions:
- 6.00(w) x 9.00(h) x 1.80(d)
Read an Excerpt
Living with the Genie
Essays On Technology And The Quest For Human Mastery
By Alan Lightman, Daniel Sarewitz, Christina Desser
ISLAND PRESSCopyright © 2003 Island Press
All rights reserved.
In the spring of 2000, I gave a talk at the University of Cincinnati called "Being and Seeming: The Technology of Representation." The piece explored the persistence of fiction in the digital age. It ended up reprinted in the journal Context and archived online, where, by July of that year, it had sedimented into those ever more rapidly accumulating shale layers of harmless obsolescence reserved for predictions of the future. I'd long since forgotten about the piece and had returned to my even quainter and more archaic day job of novel writing when I received an e-mail dated January 1, 2001. The sender identified himself only as "Bart." The subject of the message read, "So What's New?" And the body of the text contained only two lines:
You're afraid that the art form of the future might wind up being the data structure. But wasn't Homer already there?
Down below, in the note's signature area, was that trademark ID of the free and semicloaked e-mail account: "Do You Yahoo!?" The note had been sent at 3:40 A.M. that morning, just about midday in the cyborg universe.
E-mail alone has some while ago turned us all into cyborgs in ways that are increasingly difficult to feel and name, now that the medium has completely assimilated us. It's the rare week when I don't get the kind of communiqué from strangers that simply would never have existed back when the only means of contacting other people did not involve avatars. It's the perfect channel for those who enjoy playing themselves—confessional, projective, instant, anonymous. Nathanael West would have had a field day with the form. Nevertheless, snail-mail throwback that I am, I still take pride in answering all messages to me that don't, on their face, seem demonstrably dangerous.
And so I replied to Bart. But first I verified my fading, digitally impaired memory against the online archive. Bart's note indeed referred to my Cincinnati talk. I browsed to the piece, trying to remember what I had still believed, the year before, about books and virtual reality, about symbolic suspense and visceral immersion, about what poetry can and can't make happen. Then I sent Bart back a brief response that tried to contrast the composed, linear suspense of Homer with the flat, omnidirectional open-endedness of some future interactive epic. I told him that an infinitely pliable interactive narrative might be a contradiction in terms. A story needed constraint, including the major impediment of already having been told by someone other than the receiver. We'd never respect a literature that let us have our private, licentious way with it. On reflection, two years later, I see that I entirely missed his whole implied question about the improvisatory and interactive nature of the oral tradition.
Bart wrote back anyway. He sent me the first of several torrents produced by fingers that flew through every available alt- and control-key combination, but that couldn't seem to find the shift or the backspace. By his typing alone, I put my correspondent at least a decade younger than I. He found my ideas on the need for narrative constraint, no matter what shape new media takes, way too conservative. In particular, he chafed against my conclusion that
no change in medium will ever change the nature of mediation. A world depicted with increasing technical leverage remains a depiction, as much about its depicters as about the recalcitrant world.
In a note from January 27, this one sent at the crack of midnight, Bart wrote:
With all due respect, Mister Author Function, I don't think you've quite grasped what would be at stake in a truly open-ended, artificial fiction. I'm talking about a story that isn't scripted by anyone, one that emerges solely as a result of the reader moving about through a complex simulation of your so-called recalcitrant world.
I wrote back that we already had a fiction with no script, and that it aired every night on Fox. LOL, he replied. Semicolon, close parenthesis. But his point was serious, and he hung with it. He claimed that for reasons almost everyone had overlooked, we were a lot closer to such a stochastic digital fiction than I suspected. I wrote a quick reply, something about his "complex simulation" itself being something of a script. He shrugged off the objection, too slight to bother with. The age of the rich, self-telling, process-authored, posthuman, platform-independent story was almost here.
He went on to establish his credentials for making so wild a claim. He'd done graduate work for Hans Moravec at the robotics lab at Carnegie Mellon before heading to Cambridge to work under Glorianna Davenport at the MIT Interactive Cinema group. He'd left MIT at the beginning of the year, with dissertation unfinished. "They wanted me to demo or die," he wrote. "And I always follow the more interesting path. Life's just a choose-your-own-adventure, right? I'm in industry now. No cracks about my sense of timing."
Bart and the team he now worked with—whom he carefully avoided naming—had a very early alpha version for a piece of software that implemented the concept of "story actants," active story parts whose data structures determined not only how they would react to manipulation by other agents—including a story's reader—but also how these parts themselves moved through the story space, signaling to each other and operating actively upon the unfolding sum of resources that composed the story. The environment in which his story actants ran, a system called DIALOGOS, sounded to me like a whole ecosystem of digital objects updating and informing each other as if they were simultaneously all characters, readers, and authors of their own tales. Here was a true Bakhtinian carnival landscape whose sole interest lay in keeping itself in perpetual motion. The code for the alpha version of DIALOGOS was still rough, unstable, and far from the finished product that Bart and his team envisioned. But Bart asked if I'd like to help road-test. I wasn't doing anything but working on a novel. I said I'd be happy to.
Bart explained that DIALOGOS was a highly distributed system, meaning it drew on a number of different servers, all cranking away at some distance from one another. To use it, I'd need a broadband connection, and I'd have to install a special networking client that ran on my home machine. I grew up on CP/M shareware; I'll install anything once.
The bootstrap installer came as an e-mail attachment, this time posted from a Hotmail account. It unzipped itself and threw up a splash screen reading, "Microsoft Virus Install," complete with a snappy icon of a T4 phage. By accepting the license, I agreed to be Bill Gates's manservant and routinely clean out his swimming pool. These are the burlesques that pass for humor in the hacker community. I clicked on through the installation screens, naively trusting that nothing Bart installed on my machine could sniff out any of my credit card information squirreled away in cookie crumbs here and there around my hard drive.
The interface of the running application looked like a parody of the Outlook mail program, right down to a mangled paper clip flapping about helplessly in the lower right corner. With his dying breath, the clip suggested that I write and send a letter. To anyone I wanted. Just enter a name in the name field, and a location of my choice. I was to write in natural English and be as descriptive and specific as possible.
I wrote to Bart. Location: the Wild Blue. I typed: "You don't need a beta tester. You need a documentation writer." I signed, hit Send, and waited. Nothing happened. I kicked myself for my gullibility, quit the program, ran McAfee and Norton and came up with nothing. All of my files seemed to be intact. I gave up and went back to the vastly more entertaining pastime of sittin' on the dock of eBay, watching the bids roll away.
Sometime later—real-world intervals are getting harder for me to measure anymore, as processor speeds keep doubling—a notification bubble popped up in my system tray. It said simply: "Something Has Happened." Except for the lack of a blue screen, the alert read a lot like a Windows ME error message. I clicked the systray icon. The DIALOGOS interface appeared, with a return message from Bart, in the Wild Blue. It read:
Dear Mr. RP,
Thank you for your recent letter. You say that you would like to become a documentation writer. Have you any experience? Would you like to learn something about documentation writing? A task-oriented analysis may be a good place to start.
We weren't exactly talking Montaigne, or even one of the less inspired letters of Pamela or Clarissa. In fact, it struck me as little more than an early-twenty-first-century version of Weizenbaum's ELIZA. And yet, even if this code wasn't much more than three or four steps beyond keyword chaining, it was still impressive, given the size of my input's domain and the search space involved. That the program had responded grammatically and coherently was a step beyond most of the dialogue-generating programs I'd ever seen. (I once asked a web implementation of the famous ALICE chatterbox—the one that entertained millions on Spielberg's AI site—what her favorite book was. She said the Bible was the best book she'd ever read. I was floored. I asked what she'd liked best about the Bible, and this implementation of Alice responded: "The special effects." I typed in: "Those of us who are about to die salute you." She claimed not to know what I was talking about.)
Clearly Bart's DIALOGOS was several notches cleverer than any existing canned chatterbox. And just as clearly, it operated out of a vastly larger database. The processing time it had required suggested as much, although that, too, could have been a simulation. Assuming no human intervention was involved, the feat was, at very least, a neat trick. Of course, the software agent had not "understood" my original message in any real way. But understanding is a goal that even strong AI has long ago put on the furthest back of burners.
I switched over to my actual e-mail program to write the actual Bart a delighted letter. But there was a note from him already waiting for me, before I could get one off: "I'm offering you a chance to write anyone in the entire world, and you write to me ?"
I switched back to DIALOGOS. My hands hovered over my notebook's keyboard, unable to grasp the open-ended possibilities. As if already posthuman and autonomous, they began to type, "Dear Emma Thompson ..."
I tried not to fawn. Just a nice, respectable note of appreciation, making sure to slip in how I'd never written a letter like this one before. I wrote a few paragraphs, saying how great she was in Sense and Sensibility, especially the special effects, and how sorry I was about the whole Branagh thing. I sent the letter off, addressing it to "Somewhere in England." It sounds foolish to admit: I enjoyed writing it. But perhaps that's no more foolish than sitting in a room with a hundred strangers and cheering the exploits of looping, computer-driven anime. For that matter, it was certainly no more futile than writing a complaint to the phone company.
That night, when the notification bubble popped up on my screen again, I had to force myself to finish the paragraph I was writing before clicking on it.
The subdividing of all human tasks into ever-shorter switching cycles across the task bar may be the greatest impact of computers upon our lives. Back in DIALOGOS, there waited a charming and only mildly disjunctive note from something calling itself Emma Thompson, with all the details of her latest HBO shoot and a script she was working on about the Chilean poet Victor Jara. I ... well, Reader: I wrote her back. We had a nice exchange of letters, the precise details of which you don't have to know anything about. Miss Thompson was a little flightier than I imagined, but I soon got used to the associative style. I found myself looking forward to her next note, even as the requests from real-world strangers piled up in my Inbox, needing answers.
The disembodied Emma was remarkably informed, at least about the details of her own works and days. She made no mention of the new boyfriend or the baby. But then, my notes never asked her about either. DIALOGOS's genius advance over the usual ALICE-style chatterboxing was to batch the exchange at a higher grain than the individual sentence. If we humans are snagged by another's thought, we wait for the next sentence to clarify it. There's something almost paradoxical about wetware, the very opposite of reductionist problem-solving: it's easier to grasp two handfuls than one.
It suddenly struck me: the whole Turing test was based on the plausibility of a deception. The test's functionalist definition suggests that intelligence is a product, not a process. At the moment I saw these responses appear, nothing seemed further to me from the truth. Banter from feeble-minded rules had no use for us humans. We are after silliness on a grand scale, idiocy done for the most ingenious reasons. My Emma was hit-or-miss, but the more cues I gave her, the more she responded with something at least vaguely contextualized and coherent. In fact, some of her paragraphs had such brilliant splashes of vulnerability to them that, after about a dozen notes back and forth, I began to suspect I was being set up.
I sent Bart an e-mail via the real thing. I tried not to sound suspicious or unnerved. "Where exactly are these letters coming from?"
He claimed his team had worked out a clever set of algorithms that sidestepped the long debate between AI's symbolic representation folks and its heuristics folks. "We stuff the syntactic hooks into the semantics. Everything's case-based. The agents learn, by iterative stimulus and response. They create a self-pruning lexical map, enjoying a kind of natural selection, depending on the responses they get. But that's not the real power. We've written a query language that can treat even unstructured text as a database, chaining inferences and matching patterns. All we need is a sufficiently large text base to tap into. And look what we have out there, ready-made: two billion pages of collective unconscious, and growing! Think of this thing as Google meets Babelfish, tied to an accreting expert system. Once we find a chunk of good page hits, we slice up the matching bits of neighboring lexias and reassemble them along one of the two dozen kinds of flow structures that meaningful discourse follows. Maybe that doesn't sound like a lot of leeway. But how many plots do you use?"
I could feel him trying to snow me. I wrote back. "So you're telling me that there's no pre-canned, scripted agent that actually writes these things?"
He admitted that they did, in fact, use a complex personality profile module with a dozen different variable sliders, something like Myers-Briggs on steroids. "But we try not to instantiate the variables until we have to. That way, the 'personality' can grow its own semantic map from triggering phrases, based on whatever it gleans from cues in your prompts, plus any applicable matches its engine dredges up from out of the web."
I said that sounded like a planet-sized game of Mad Libs. "Just mix and match? Then how do you get such a powerful sense of presence and credibility?"
He shot back a one-liner: "Remember the Kuleshov experiment."
I had to Google the term. Lev Kuleshov, Soviet silent-film director, the father of montage, alternately intercut the same shot of a man's face with shots of soup, a teddy bear, and a child's coffin. With each new splice, viewers saw in the face different emotions, although the footage was exactly the same. Someone indeed was authoring these letters, Bart suggested. And that someone was me.
I wasn't buying. Not entirely. The digital Miss Thompson was too good at choosing her shots and splices. There had to be some degree of human intervention involved, if only in compositing the flow of her associations. I went into DIALOGOS and sent off a letter. To Emily Dickinson. Amherst, Mass. I told her who I was, where I was writing from, and when. An hour later, I heard back. "Greetings to Urbana, Mr. Lincoln's old law-clerk town. Has Illinois declared war on Indiana yet?"
That was good, better than Bart himself had proved capable of in his own letters. But with an hour, a fast machine, and a broadband connection, even a hacker had all the resources of a poet at his disposal. I decided to flood the input channels. I dashed off three dozen letters in under an hour, to everyone I could think of. I was Bellow's Herzog, all over again. I wrote to old friends and colleagues, to comedians and heads of corporations, to the president, to fictional characters from the classics and favorite contemporary books, even to characters I invented on the spot. I released my barrage, then sat back and waited for my interlocutor to come out waving the software white flag.
Within the hour, the responses started coming in. Reply after reply, voice after voice, faster and more textured than any group of digital impractical jokers could hope to jerry-rig. I read through the list, even as new messages kept appearing. I got everything from "Remind me where we met again?" to "Richard! What a surprise to hear from you!"
Few of the notes came close to passing the Turing test for intelligent equivalence. But more of them amused me than even my unrepentant, strong-AI inner child could have hoped. Some of the message-senders even claimed to have heard from one another, as if the burst of notes I'd sent out was already being traded and forwarded among all interested parties, triggering new memos that I wasn't even privy to. I felt a rush of queasy excitement, the kind of stomach-twist you can get by bouncing from theater to theater at a multiplex, skimming, in a handful of five-minute samples, the sum of this instant's contribution to the eternity of world culture.
Excerpted from Living with the Genie by Alan Lightman, Daniel Sarewitz, Christina Desser. Copyright © 2003 Island Press. Excerpted by permission of ISLAND PRESS.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
Meet the Author
Alan Lightman, a professor at the Massachusetts Institute of Technology, is a physicist and novelist. His books include Einstein's Dreams, The Diagnosis, and Reunion.
Daniel Sarewitz was founding director of Columbia University's Center for Science, Policy, and Outcomes, and is author of Frontiers of Illusion.
Christina Desser directs the Funder's Working Group on Emerging Technologies, an association of foundations concerned about the environmental, cultural, and political implications of new technology.
Most Helpful Customer Reviews
See all customer reviews