Nick Bostrom, a professor at Oxford, argues we might be living in a created simulation. His logic entails that we should conclude right now that it’s about 20 percent likely we live in a simulated world. Now, this possibility may be irrelevant since we’re far away from creating a simulated world.
Subjective and Objective Probability
Nick Bostrom is talking about subjective probability here, not objective. Objective probability is the kind of probability that people use when they are talking about throwing a die. Before it is rolled, there is a one in six objective probability that a three will be rolled on a six-sided die. However, once it has been rolled, the objective probability goes to either 100 percent or zero percent because it either comes up with a three or it doesn’t.
Subjective probabilities, however, are expressions of how likely something is, given what we know—given a set of evidence. The subjective probability of a three before the dice is rolled is the same: it’s one in six.
But suppose the dice is rolled and then covered with a cup. The objective probability is still 100 percent or zero percent because it either came up a three or it didn’t.
But subjectively, given what is known, ‘three’ should still be assigned a probability of one in six because that is how likely it is, given what is known.
So, objectively, the probability that we are in a simulated world is either 100 percent or zero percent—because either we are or we are not. But the subjective probability will turn on what we know or have good reason to suspect.
This is a transcript from the video series Sci-Phi: Science Fiction as Philosophy. Watch it now, on Wondrium.
The Five Predictions of the Future
The subjective probability assigned to us living in a computer simulation should be directly proportional to how likely it is that humanity will one day create simulated worlds. And Bostrom thinks that this is about 20 percent likely. Thinking about whether we will create simulated worlds, there are essentially five possibilities.
First, we may simply blow ourselves up or destroy our civilization in some other way before we ever get that technologically advanced. Maybe climate change caused by global warming will make the species go extinct. So that’s one reason it may never happen.
Second, ethical considerations could prevent us from creating them. Simulated people would be sentient, so they would actually feel pain and have rights. It would be wrong to force them, for example, to live in a world in which Hitler won World War II just to satisfy our curiosity.
Third, we might lose interest. Maybe we discover that directly stimulating the pleasure centers of our brain is better entertainment, become obsessed with that, and stop advancing scientifically.
Fourth, there could be technological limitations that just make simulated worlds impossible. Maybe it’s impossible to simulate what a brain does with microchips.
And fifth, there is the possibility that we actually could and will develop computer-simulated worlds.
Learn more about Transcendence and the dangers of AI.
But Why 20 Percent?
Now, given that we don’t know anything, each scenario is equally likely. So we should assign each one an equal probability. One hundred divided by five is 20. So it’s 20 percent likely that we will lose interest, it’s 20 percent likely that it’s technologically impossible—but it’s also 20 percent likely that we will, one day, create simulated worlds.
And if we do, because of logical reasons, it’s a near guarantee that we are in a simulated world. Although the way subjective probability works is that the likelihoods that are assigned are updated when new information comes in.
So, say, the evidence shows that we have already dumped so much carbon dioxide in the atmosphere that runaway global warming is inevitable, and the planet will become uninhabitable. That would greatly lower the subjective probability that we will develop simulated worlds, and thus the subjective probability that we are in one.
Learn more about The Orville, Orwell, and the Black Mirror.
Quantum Mechanics as Evidence of Us Living in a Created Simulation
Some have objected to Bostrom’s argument that the computing power to simulate the entire universe, down to every atom, is just unfeasible. But a simulated world would not have to bother with all that.
It would not have to simulate every atom in a tree when it is looked at; it would just send a ‘see a tree’ signal to the brain. Only if we started looking at things on the atomic level would it have to bother to render anything on that level, and only when and where it was being observed.
The computer would not need to assign a specific location or momentum to electrons; an equation to keep track of where they would likely be when they were measured would be good enough. It could use that to simply render them on the occasions that we happen to take a measurement.
Individual particles don’t have location or momentum. They are wave functions—equations—until they are measured. And wouldn’t the fact we are in a simulated world explain perfectly why they don’t collapse until they are measured? The simulation knows we are measuring and thus knows it needs to collapse the wave function.
Quantum mechanics doesn’t prove that we are in a computer simulation. But since the simulation hypothesis would explain quantum mechanics so well—it explains what a wave function is, it accounts for the measurement problem—maybe thinking ‘we live in a computer simulation’ should be considered a viable interpretation of quantum mechanics.
Common Questions about the Probability We Live in a Created Simulation
Our civilization might be destroyed before we get the chance to build a created simulation. Or we might not build one because of ethical reasons. There’s also the chance of us losing interest in simulations or there being technological limitations making simulated worlds impossible.
Yes, since it’s a subjective probability, it changes with new information. So the probability of us living in a created simulation might change based on information gathered in the future, like being assured that global warming has gone too far.
Since the simulation hypothesis, that only objects that needed to be seen would be created, would explain quantum mechanics so well, maybe thinking ‘we live in a computer simulation’ should be considered a viable interpretation of quantum mechanics.