Sixth in an ongoing series about the places where science and religion meet. The previous episode is here.
The sensory aspect of creating a convincing virtual-reality video game seems like a surmountable technical challenge, and the insertion of the real-world player into the game-world avoids the hard problem of consciousness. But video-game worlds are typically well-stocked with "non-player characters" (NPCs), computer-generated entities who interact with the player, sometimes as allies, and sometimes as adversaries. Is it possible, as philosopher David Chalmers believes, that these could be “philosophical zombies,” creatures that give all the appearances of consciousness, but without actually possessing it?
In current-day video games, the believability of NPCs is strictly limited. Some artificial intelligences can seemingly exude wit and personality, but only when given the appropriate, preprogrammed cue. There is no actual interaction taking place, just the mechanical regurgitation of a specific, anticipated response. This can lead to surreal conversations, as exemplified by digital assistant Siri’s often loopy misunderstandings of simple questions (or the repetitious habit of multiple characters in the video game The Elder Scrolls: Skyrim to remark, “I used to be an adventurer like you, until I took an arrow in the knee.”). This kind of emulation of intelligence is eerie and impressive when the programmer correctly anticipates your input, but quickly reveals itself to be as flimsy a sham as the children’s toys that mimic conversation by simply repeating back the child’s own words.
Although there have been attempts at creating true simulations of intelligence, machines that can learn and respond appropriately to unbounded input, they have not, as of the time of this writing, progressed significantly far in the way of believably duplicating human interactions (although they have mastered tasks as diverse as as playing chess, competing on television game show Jeopardy, and identifying other robots as robots). Are these major steps on the pathway, or deceptive dead ends? Could technology ever improve to the point where it could convincingly simulate, not you perhaps, but other people, in all their deep, multifaceted, and endlessly surprising soulfulness? Is true artificial intelligence, to the point that computers could believably create people, actually achievable?
As with the hard problem of consciousness, it is possible to sidestep this problem by selectively importing resources from the real world. Many current video games are what is known as “massively multiplayer”; they feature what is sometimes up to millions of real people all interacting in a single continuous, unified game-world. So if our world was a multiplayer video game and not a single-player experience, it's possible that some people in our world could be real and some could be fake. The ones we interact with all the time—close friends, family members, and so forth—would be real, the ones we just see from a distance, or have minor interactions with, or read about in the news might be simulated.
There are a few problems with this idea, however. First of all, we all have different “circles” of friends and associates in different places. If we're assuming that each real person interacts largely with other real people, and that all the places real people go must be fully and believably rendered by the game engine, then it does not take long before the game world must be actually as large and as fully explorable as it seems to be (although still without the unimaginable vastness of outer space). That in turn might reasonably be expected to tax the resources of the video-game designers and the computer they're working with.
A bigger problem is more philosophical: what would be the point of it? At this point, we're describing a fake world that is essentially copied from and populated by the residents of a real world very much like it. Why would vast numbers of people choose to leave that world and enter this one? What would be the advantage or appeal?
Of course, we could picture a paranoid, Matrix-inspired scenario, in which our world is a mental prison, built to keep us distracted and mentally occupied while nefarious forces are at work. Or conversely, the world really could be a one-player video game, but most of us inside it could have been impressed into service to be the minds animating the NPCs. (Even worse, we could actually have been bred from birth to be NPCs, a possibility that eliminates the need for induced amnesia and false memories.) Perhaps the real future world is postapocalyptic and bleak, and people are trying to re-create their former lives. Maybe their real world is actually quite different from ours, and what is mundane reality to us is fantastical escapism for them. Or maybe we are living the Bostrom scenario, where our future descendants are merely curious about what living like their ancestors would feel like.
The problem here is that we have entered the typical science-fiction territory of plausible departures from reality-as-we-know-it without any compelling grounds for belief, or reason to choose one scenario over another. We could all be sleeping in Fairyland and dreaming this world, or living in a naturalistic habitat in an alien zoo somewhere. There is as much or as little reason to believe those fantasies as any of the others detailed above, and one very good reason to disbelieve all of them: They needlessly complicate our ontology because they require unsubstantiated, undermotivated belief in an additional level of reality that is presumably at least as vast and complex as our own.
This brings us back to where we began this series, the simulation theory of Professor Nick Bostrom. His hypothesis, although requiring sharper departures from current technology, was allied to a chain of reasoning that claims we have a compelling reason to suspect it to be true. But does his argument actually hold water?
Barrat, James, Our Final Invention: Artificial Intelligence and the End of the Human Era, Thomas Dunne Books, New York, 2013.