Elon Musk, the genius behind SpaceX and Tesla, surprised the California Code Conference by asserting that we’re probably living inside a computer simulation. While this sounds unlikely to the untrained ear, Musk’s business ventures are rooted firmly in cutting-edge technology, and his bold suggestion is rooted in the hard science he works with every day.
Musk points out that our ability to create realistic and immersive video games has evolved rapidly over the last 40 years. Within a few years, he says, we will have technology advanced enough to create virtual worlds that are both indistinguishable from and seamlessly merged with the world we interact with every day. This melding of the physical world with computer-generated content is called augmented reality (AR). Combining AR with artificial intelligence (AI), an all-but-certain outcome of our current technological research, will allow us to form artificial realities that are indistinguishable from the real world.
To lightly paraphrase Musk’s argument, “If we grant any rate of improvement at all, games will soon be indistinguishable from reality. That’s true even if the rate of advancement drops to a thousandth of what it is now. Now, imagine what technology will look like in 10,000 years, which is nothing on the evolutionary scale.
“No matter when it happens, we will have games that are indistinguishable from reality. Those games could be played on a console or on a PC or whatever, and there would be billions of people would have these computers or consoles. That would mean that there are billions of hyper-realistic, man-made worlds in the future, each with characters powered by genuine AI. If all these worlds are in one base reality, the odds we’re in that base reality is one in billions.”
Musk isn’t the first to suspect that Simulation Creationism is the source of our reality. The concept that our world is externally constructed, with or without using computers, has been suggested by philosophers and physicists alike for centuries. In fact, Nick Bostrom came to the conclusion that we probably live in a simulation in 2003.
Musk agree on more than the idea that we live in a computer simulation, however. Both also believe that AI could prove very dangerous. Musk believes that developing true AI will end civilisation as we know it. Bostrom foresees a similar outcome if developers don’t assess and account for the risks appropriately.
Fact or Fiction?
How likely are these men to be correct? The idea of an AI-driven apocalypse and a computer-simulated world are recurring themes in science fiction, as can be seen in The Matrix and 2001: A Space Odyssey, but are Musk and Bostrom drawing our attention to events that could really occur?
Resource arguments strongly suggest that we aren’t living in a simulation. The computing power necessary to run a perfect simulation of real life would be astronomical. Such a system would need to keep track of every being in the world, along with all of their interactions. That alone would require immense processing power, and that doesn’t even account for the difficulties of rendering static objects and the laws of the universe. Quantum mechanics also suggests that we live in the real world, as it supports the idea that running a city-sized simulation would require a city-sized computer. These arguments make it seem unlikely that we live within a simulation.
If we built a machine that could simulate our lives there is a good chance the simulation would contain glitches we would encounter. Simulating a world as complex as ours would invite any number of bugs into the system. For example, stars may only be visible at certain levels of magnification. We’ve never noticed these sorts of glitches, despite the world having been subject to intense scientific scrutiny for decades.
As for AI, machine learning remains clunky and fails to reach a level of cognition that comes close to what humans can do. Genuinely intelligent software isn’t even on the horizon, and will take many years to create. Currently, the best we can do is meticulously program software to perform well in specific contexts. Machines that can take over the world are a very distant nightmare.
One of the better forms of AI we currently possess are called neural networks, which are electronic models of the brain. We can use them to predict stock price changes. Neural networks for stock prediction are trained with historical stock trading information to help them identify patterns in live streams of stock data. They use the patterns they recognize to predict outcomes, and traders use this information to maximize profits and minimize losses.
We also have AI that uses knowledge-based systems. These are programmed with many rules and facts to make doing jobs easier and faster for people. People who use these systems may or may not know they’re interacting with a machine like Jill, the online tutor of AI, which helps AI students by answering questions and providing feedback. Other knowledge-based AI programs are developed to work on or within specific domains or to solve a narrow set of problems.
Given that AI necessarily masters a very small area of knowledge, and AI with broad-based, flexible knowledge has yet to be developed, the danger of AI destroying our civilization seems miniscule. As it stands, AI cannot even replace human decision-making, it can only augment it.
However, some of Musk’s predictions seem significantly more likely.
One is the broad adoption of augmented reality. Computers and smartphones alike have become a daily fixture of most of our lives, and the amount of our lives that are spent connected to and using these devices is steadily increasing, as are the ways in which these devices impact our everyday actions. We continually want more access to information and easier, quicker communication with others. This drive for increased technological reliance has encouraged the development of wearable devices.
Musk believes that we’ll ultimately become the minions of our AI if we don’t develop quality brain-computer interfaces. However, Steve Mann, who initially developed AR and wearable devices, believes we can combine both technologies to help society advance. Mann focuses primarily on assistive medical systems, which have a clear benefit for all of us. One of his ideas is to create brain implants that use electrical signals to allow movement in paralysed limbs.
All told, we’re probably not living in a computer simulation, and Musk’s claims are unlikely to come to pass. He does, however, hint at some future technological developments.
AR, AI, and associated developments will allow us to live in a more connected world. With augmented reality, we will have instant access to data and benefit from digital imagery superimposed upon what we see around us. AI will help us understand information, allowing us to make better decisions. But for all AI and AR can and will make our lives better, their impact will exist in the real world.