conducted by: GF Willmetts. In the past five years, I’ve interviewed Robert Sawyer a couple times, so this makes for the hat-trick. After the single season of ‘FlashForward’, his name is now known globally so there is a lot more to chat about. Rob describes my interviews as a pleasant torture, which also means he’s a glutton for punishment, so let’s adjust the rack I’ve tied him to and get down to some serious questioning.

SFCrowsnest: Well, Rob, a couple years since our last interview and because of the TV series ‘FlashForward’, based on your book, the whole world knows your name now. So, despite not getting to a second season, what's the Hollywood treatment like?

Robert J. Sawyer: It was a fabulous experience. I was treated well and with respect, consulted on every episode, wrote the 19th episode (‘Course Correction’), made lots of friends with the writers, cast and crew, was paid well initially and continue to receive fat residual payments. It was one of the best experiences of my life.

SFC: When we were building up for this interview, you said you were kinda interviewed out on the subject of ‘FlashForward’. For the record, folks, I did know Rob was really busy this past year and I doubt if he could talk about the series even if he was allowed to. I presume you had some idea where the series was going?


Photo Credit: Carolyn Clink
RJS: I was in the writers' room when we were hacking out how the first season might end. But the reality is that television isn't created in stone; it's modified as the weeks go by, there are behind-the-scenes staff changes and there are ever-changing demands from the network and the studio. Unlike the visions of the future in my novel - which showed a fixed, unchangeable reality - the future of the TV series was a work in progress and, sadly, one that never got finished.

SFC: Does this mean no one knew where the final ever episode was going to end up had it got to the end of the series?

RJS: Well, there were several strongly held opinions about how it should end, but television is a collaborative medium: no one, for any series, knows how it's going to end years in advance - or who is going to be in the position of making those choices when the end finally comes. That's the nature of the medium. On the other hand, I knew exactly how my trilogy 'Wake', 'Watch' and 'Wonder' was going to end; I wrote the epilogue first, six years before I finished the series and it stayed pretty much verbatim the same in the final product. And THAT'S the nature of the different medium of writing books with a single author.

SFC: I spotted you wrote the script for the nineteenth episode. Was this to get insight into scripting or have you done this before?

RJS: My university education is in radio and television arts and I've been selling TV scripts professionally for well over a decade, as well as writing TV series bibles and so on. I wrote the bible for the TV series ‘Charlie Jade’, recently seen in the US on Syfy and I did conceptual work on ‘Robotech: The Shadow Chronicles’, among other things. I wrote the episode because I wanted an active role in the adaptation of my novel and because it gave me a further financial stake in the show. I had the only solo scriptwriting credit on the 22-episode series, so that means, in addition to the royalty I got on every episode, I also get 1/22nd - 4.5% - of the writing royalties for the entire series.

SFC: How did it feel playing with a modified version of your ideas and characters?

RJS: Oh, I had no problem with that. It'd been eleven years since I wrote ‘FlashForward’ the novel. David Goyer, the showrunner, kept saying things to me like, ‘Remember that part in your book when ...’ and I'd reply, ‘Really? I'd forgotten that.’ Books and TV are different media; I fully understand that, and never once pushed for something to be done a certain way in show simply because that's how I'd handled it in the book.

SFC: Is there anything you can tell us about the new Hollywood project?

RJS: I've just signed a very nice deal to develop a TV series based on ‘Wake’, ‘Watch’ and ‘Wonder’. I'm taking an even more active role in that, writing the pilot, executive producing, and showrunning and doing a larger number of the scripts.

SFC: I presume with the latest trilogy making a transition to TV, it's going to be a closer adaptation or at least initially because you could always extend the original trilogy as to what would happen next?

RJS: I'm one of the executive producers on the WWW TV adaptation and am attached to write the pilot script, so, yes, it's fair to say it'll hew reasonably close to the source novels.

SFC: Reading your latest series, the WWW trilogy, I was struck by how much research you must have done on a variety of subjects. Is that specific for all your books or are you picking up information all the time and just draw it out of the melting pot when you think it will fit in a story.

RJS: I do months and months of specific research for every single book; of course, I don't forget what I learned for one book as soon as I start another. For instance, I did a lot of research on consciousness for ‘Hominds’ and ‘Mindscan’, and I certainly drew upon that for the WWW trilogy, in addition to all the fresh research I did for those books.

SFC: Where are your key sources for research? Books, magazines, the Net or a combination of all three?

RJS: It's more than just those three sources: I spend a lot of time talking face-to-face real scientists. For instance, for the WWW trilogy I went to the AI lab at MIT, and spoke to AI pioneer Marvin Minsky. I also led a brainstorming session about web consciousness at the Googleplex - the international headquarters of Google - and Google co-founder Larry Page attended that. But, sure, like everyone these days, web research is my number-one source, supplemented by books, magazines, journals and attending scientific conferences.

SFC: Are there any subjects you find harder to get your head around than others, especially in explaining them?

RJS: Mathematics is always hard to explain in prose and there are readers who get turned off the moment you present an equation or a formula in fiction. It was hard to explain Zipf plots and Shannon entropy in ‘Wake’ without resorting to diagrams or equations, but I think I pulled it off.

SFC: Did you ever think you might have been going in over your head as to the Webmind's capabilities or could have gone further?

RJS: I spent six years on this project. My goal was to produce the most accurate and realistic fictional portrayal to date of a nascent AI, without any hand-waving and without the keys moments occurring off-screen. It's very hard to do that, but that's what made it worth doing.

SFC: Do you think it would benefit our own reality to have an AI on the Internet as you have in your story?

RJS: Of course. That's why I wrote the book.

SFC: I did find it odd that the first choice of the film ‘WarGames’ to watch with Webmind an unusual choice, more so that it could end up teaching Webmind how to create nuclear chaos. It's a good thing that Webmind could tell the difference between reality and fiction or it might have thought that that humans had a thing against rampant AIs.

RJS: You're misremembering both the novel and the film; in the novel, Caitlin and her father are watching a film for their pleasure on Caitlin's birthday, and the film is not about how to create nuclear chaos - no mechanism is provided for that except taking humans out of the loop in launching missiles and that's not the reality of our world - but rather about the folly of nuclear war. Also, if one reads the novels carefully, you'll see that ‘WarGames’ plays into the game-theory subtext of the trilogy.

SFC: I think I might have been giving too much ability for the Webmind to imagine what capabilities it had at its disposal knowing that it could infiltrate US security rather than the plot of the film itself. I suspect the hardest thing to educate any nascent AI about would be the difference between fiction and reality, especially some of the more, shall we say, gray areas.

RJS: Oh, not at all, fiction and reality are an easy concept. Any chess-playing computer knows the difference: here are all the possible moves that my opponent might make next - ah, and here's the one it actually chose to make: some things that might plausibly happen really do, and others don't - they're just possibilities or fanciful explorations. But 'WarGames' is a film about peace, not war - and it makes, as I said, an interesting point, about which games are worth playing, which is why I chose to spotlight it.

SFC: In many respects, Webmind is the antithesis of D.F. Jones’ ‘The Forbin Project's book AI, Colossus. Considering the number of films Webmind was introduced to, I'm surprised he didn't take lessons from the film version, even if it was something to avoid.

RJS: He's introduced to three films, which isn't really an enormous number: ‘WarGames’, ‘Star Trek: The Motion Picture’ and ‘2001: A Space Odyssey’. That said, Colossus is mentioned in Chapter 22 of ‘Watch’, but dwelling on a film that most people have forgotten would have just bogged down the narrative.

SFC: You sure 'The Forbin Project' is that forgotten? I mean, it did have a second special edition release a couple years ago. It isn't like all the SF community has forgotten about it.

RJS: But my books are only in part written for the SF community - they're mainstream bestsellers in Canada, and so I choose pop-culture references that will resonate with a large audience. Besides, as I said, ‘The Forbin Project’ IS mentioned in my trilogy.

SFC: I'm a bit puzzled how Webmind can make sense of what it can read. After all, this is a real stumbling block in creating AIs at the moment, but then having Kuroda creating algorithms for it to perceive pictures and movies was pushing it a little as well. I mean, it's a complicated process to have understanding of what you're looking at not to mention taking up computer space.

RJS: Honestly, this is covered at great length in ‘Wake’, you should go re-read that. Whole chapters are devoted to it, including discussions of how Webmind learns to recognize the visual forms of the alphabet, how it determines the Unicode sequences that correspond to each character, how it comes to access online dictionaries and its absorbing of the Cyc database, which explains in mathematical language based on second-order predicate calculus most of the background knowledge one would need to understand English; Cyc really exists, as do all the other tools Webmind accessed. And I totally disagree about the codecs that Kuroda creates; again, this material is discussed in depth in the novels. The reason it's difficult to teach a dumb machine to read or see is that the machine is dumb - utterly devoid of consciousness and utterly incapable of giving feedback; what you describe is the difficulty of teaching a pebble to read or see; Webmind is an emergent consciousness and the process is much more akin to teaching a child to read or see, something even non-specialists have great success at.

Q1a:

SFC: Do you think we should enlighten readers as to the source of the name 'Koruda', not to mention they can see the source on DVD in June?

RJS: Ah, yes! Dr. Koruda - who is a vision researcher and information researcher - in my novels is named after a character in the 1972 American TV series ‘Search’. The Kuroda in that show was a visual systems and telemetry specialist working in a high-tech mission-control room. The pilot movie for that series, ‘Probe’, was just released on DVD as part of the Warner Archive Collection.

SFC: Wouldn't Webmind only be as smart as the sum of human knowledge and inaccuracy on the Internet?

RJS: Well, first, you're being dismissive: the sum of human knowledge is a vastly wonderful thing and Webmind manages to digest it all, meaning it has more knowledge than any one person will ever have and so can make connections that have eluded us - as he does when curing cancer in ‘Wonder’. Second, sure, there's lots of inaccurate information on the Internet, but there's also lots of accurate information on the same topics and it's easy enough for you and I to distinguish between the two. Example: did we or did we not land on the moon in 1969? Any entity that has read the entire Internet will know beyond a shadow of a doubt that we absolutely did.

SFC: The problem for any AI is differentiating between truth, wind-ups, lies and misinformation. Even cross-connecting information on the Net, if enough people say the same things as misinformation, people believe it as truth. An AI is bound to have similar problems. All I'm pointing out is that it wouldn't be a perfect match.

RJS: I disagree, Geoff, but you're entitled to your opinion. Seriously, take a proposition: the Earth is billions of years old. Are you telling me that you couldn't prove that to be the consensus of informed, scientific thought by searching the Web? It's only when you CAN'T check the source - who the person is, what else they've said, what others have said in response to them - that lies can be taken as truth, which is why cults isolate people they're trying to recruit from family and friends. An infinitely patient entity with access to all that's on the web would have no trouble discerning the actual truth about almost anything that actually matters.

Q2a:

SFC: Similar to the Turing Test, where do you think the distinction between copying the creativity of others and being creative itself would develop in AIs?

RJS: Well, most of what we do is copying the creativity of others and yet we don't dispute human intelligence. I didn't invent the English language. I didn't invent grammar. I didn't invent the standard plot structure used in most novels. I didn't invent the basic constituents of drama - but I, and everyone else who writes a new novel, is being creative. We build on what has gone before. The Turing Test determines if something has reached human-level intelligence and most humans actually don't create novel works, in both senses of that word. But I don't see any barrier to artificial intelligences doing so and the test will be similar: when a novel or a painting or a song written by a computer is judged by humans who are unaware of its origins to be a human work, the AI will have passed a significant threshold.

SFC: I remember the original 'Avengers' episode, 'Love All', where a story could be created by just mixing story elements using a piano as a keyboard. I think reality has caught up with that for an AI to contrive a story. After all, a new story is only new because we haven't seen it before and few people recognise story patterns.
Art is a different matter, though. I doubt if a computer generated, AI or otherwise, would have any problems creating an acceptable surreal picture far easier than creating a portrait without going over a photograph as some art software can do. The real problem is what do we call creativity really? I suspect the critics can be fooled.

RJS: Of course they can be fooled. No question. It's been shown experimentally time and again. That doesn't mean that creativity doesn't exist. It's just hard to define - and, like so many things, one knows it when one sees it. And it is true that repetitive tasks are more easily automated than are tasks in which novelty is the desired outcome. Machines aren't forever barred from being creative, but it'll be the last thing they master and so the thing about us that they prize the most.

SFC: I was going to ask whether anyone would compare the Webmind to God until I read the third book and discovered you covered this. Do you think it's innate in humans that there is a belief that a greater being will arrive and sort out all their troubles?

RJS: It's clearly neurologically true that human beings are hardwired to see patterns that aren't there. If you think you see a tiger in the tall grass and you're wrong nine times out of ten, you're still alive; if you don't see a tiger in the tall grass and you're wrong even once, you're dead. But your leap - from an innate predisposition to belief in things that demonstrably aren't there to the contention that it's innate that we believe in something that will sort out all our troubles - simply isn't born out by history; most religions have been about vengeful or indifferent Gods. I spent much of my Neanderthal trilogy (‘Hominids’, ‘Humans’ and ‘Hybrids’) talking about what humans would be like if we'd never been predisposed to religious belief and I make the case that we'd be better people. And there's nothing whatsoever supernatural about Webmind; he's not a deity - he did not create us and did not ask us to worship him, he exists entirely in the material world and his nature is utterly explicable and comprehensible. He's not a God, but he does spur us toward personal accountability and that's all to the good, I think.

SFC: I didn't mean Webmind was a god, just how it would be perceived by some people. I mean, making the connections to curing various cancers would look miraculous to, say, the more religiously inclined who might contemplate the AI having a soul.

RJS: Which is why Webmind specifically, repeatedly and unequivocally disclaims being God. Yes, some people will always believe things in defiance of the evidence, but, really, it's a small number of people today who say an iPhone works because of magic smoke contained in the case. Most of the rest, no matter how magical an iPhone seems - in the Clarkean sense of any sufficiently advanced technology is indistinguishable from magic - recognise that it's the product of clever engineers and scientists who are mortal human beings. As to whether Webmind has a soul, you tell me what a soul is and I'll answer that - and if your answer is a sense of self that survives destruction of the physical body, then I DO answer that definitively in 'Wonder'.

SFC: Do you think that there is such a thing as the Pandora Protocol to eliminate nascent Artificial Intelligences where they occur in case they go rogue in our reality?

RJS: Certainly. Governments spend all kinds of money preparing contingency plans and reports. Heck, I'm off in September to consult with DARPA on their 100 Year Starship Project - which I think is a wonderfully valuable initiative, but also is indicative of the fact that government agencies do spend a lot of time blue-skying about things that aren't immediate concerns. And I'm glad they do, because when AI does emerge we won't have to time to strike a committee and think about it then.

SFC: I did ponder on how ineffective the Webmind would be without human helpers but then I thought that if it can access the banking system it could equally cause even more chaos. Did you consider how far you could go or have a clear idea where the trilogy would end?

RJS: I knew exactly how the trilogy was going to end; the one sample chapter I submitted with the proposal that sold all three books was the epilogue of the third volume, ‘Wonder’. And, gently, you're missing the point if you're contemplating Webmind without human helpers: the entire trilogy is about the potential win-win relationship between humans and AI, finding a way for both to prosper and us to retain our essential liberty, dignity, and individuality.

SFC: I suspect I've just seen too many unhappy ending Science Fiction stories. Generally speaking, Rob, would you agree that your stories tend towards being positive than negative?

RJS: Oh, absolutely. In part that's because it's my natural disposition and in part, because it's actually creatively more interesting and challenging to write positive, even Utopian, visions that nonetheless are exciting to read about than it is to write dystopian visions.

SFC: Have you ever contemplated using really negative or bad consequences when balancing your stories, even if you sort it out at the end?

RJS: Of course I have. Mary Vaughan is raped in 'Hominids', Tom Jericho has terminal cancer in 'Calculating God', all of Earth is destroyed in 'Golden Fleece'. If you want to see me at my nastiest, read my story ‘On the Surface’, in which the Morlocks of H.G. Wells's 'The Time Machine' mass-produce time machines and go on a rampage.

SFC: Did the thumbscrews hurt too much this time? :-)

RJS: Not at all. Did the answers? ;)

SFC: Nope, but I’m sure I can tighten the rack next time. Thanks for the interview.

© SFCrowsnest and Robert J. Sawyer 2011

All rights reserved