Макс Тегмарк - Life 3.0 - Being Human in the Age of Artificial Intelligence

Although the majority of interactions occur in virtual environments for convenience and speed, many minds still enjoy interactions and activities using physical bodies as well. For example, uploaded versions of Hans Moravec, Ray Kurzweil and Larry Page have a tradition of taking turns creating virtual realities and then exploring them together, but once in a while, they also enjoy flying together in the real world, embodied in avian winged robots. Some of the robots that roam the streets, skies and lakes of the mixed zones are similarly controlled by uploaded and augmented humans, who choose to embody themselves in the mixed zones because they enjoy being around humans and each other.

In the human-only zones, in contrast, machines with human-level general intelligence or above are banned, as are technologically enhanced biological organisms. Here, life isn’t dramatically different from today, except that it’s more affluent and convenient: poverty has been mostly eliminated, and cures are available for most of today’s diseases. The small fraction of humans who have opted to live in these zones effectively exist on a lower and more limited plane of awareness from everyone else, and have limited understanding of what their more intelligent fellow minds are doing in the other zones. However, many of them are quite happy with their lives.

AI Economics

The vast majority of all computations take place in the machine zones, which are mostly owned by the many competing superintelligent AIs that live there. By virtue of their superior intelligence and technology, no other entities can challenge their power. These AIs have agreed to cooperate and coordinate with each other under a libertarian governance system that has no rules except protection of private property. These property rights extend to all intelligent entities, including humans, and explain how the human-only zones came to exist. Early on, groups of humans banded together and decided that, in their zones, it was forbidden to sell property to non-humans.

Because of their technology, the superintelligent AIs have ended up richer than these humans by a factor much larger than that by which Bill Gates is richer than a homeless beggar. However, people in the human-only zones are still materially better off than most people today: their economy is rather decoupled from that of the machines, so the presence of the machines elsewhere has little effect on them except for the occasional useful technologies that they can understand and reproduce for themselves—much as the Amish and various technology-relinquishing native tribes today have standards of living at least as good as they had in old times. It doesn’t matter that the humans have nothing to sell that the machines need, since the machines need nothing in return.

In the mixed sectors, the wealth difference between AIs and humans is more noticeable, resulting in land (the only human-owned product that the machines want to buy) being astronomically expensive compared to other products. Most humans who owned land therefore ended up selling a small fraction of it to AIs in return for guaranteed basic income for them and their offspring/uploads in perpetuity. This liberated them from the need to work, and freed them up to enjoy the amazing abundance of cheap machine-produced goods and services, in both physical and virtual reality. As far as the machines are concerned, the mixed zones are mainly for play rather than for work.

Why This May Never Happen

Before getting too excited about adventures we may have as cyborgs or uploads, let’s consider some reasons why this scenario might never happen. First of all, there are two possible routes to enhanced humans (cyborgs and uploads):

1. We figure out how to create them ourselves.

2. We build superintelligent machines that figure it out for us.

If route 1 comes through first, it could naturally lead to a world teeming with cyborgs and uploads. However, as we discussed in the last chapter, most AI researchers think that the opposite is more likely, with enhanced or digital brains being more difficult to build than clean-slate superhuman AGIs—just as mechanical birds turned out to be harder to build than airplanes. After strong machine AI is built, it’s not obvious that cyborgs or uploads will ever be made. If the Neanderthals had had another 100,000 years to evolve and get smarter, things might have turned out great for them—but Homo sapiens never gave them that much time.

Second, even if this scenario with cyborgs and uploads did come about, it’s not clear that it would be stable and last. Why should the power balance between multiple superintelligences remain stable for millennia, rather than the AIs merging or the smartest one taking over? Moreover, why should the machines choose to respect human property rights and keep humans around, given that they don’t need humans for anything and can do all human work better and cheaper themselves? Ray Kurzweil speculates that natural and enhanced humans will be protected from extermination because “humans are respected by AIs for giving rise to the machines.”1 However, as we’ll discuss in chapter 7, we must not fall into the trap of anthropomorphizing AIs and assume that they have human-like emotions of gratitude. Indeed, though we humans are imbued with a propensity toward gratitude, we don’t show enough gratitude to our intellectual creator (our DNA) to abstain from thwarting its goals by using birth control.

Even if we buy the assumption that the AIs will opt to respect human property rights, they can gradually get much of our land in other ways, by using some of their superintelligent persuasion powers that we explored in the last chapter to persuade humans to sell some land for a life in luxury. In human-only sectors, they could entice humans to launch political campaigns for allowing land sales. After all, even die-hard bio-Luddites may want to sell some land to save the life of an ill child or to gain immortality. If the humans are educated, entertained and busy, falling birthrates may even shrink their population sizes without machine meddling, as is currently happening in Japan and Germany. This could drive humans extinct in just a few millennia.

Downsides

For some of their most ardent supporters, cyborgs and uploads hold a promise of techno-bliss and life extension for all. Indeed, the prospect of getting uploaded in the future has motivated over a hundred people to have their brains posthumously frozen by the Arizona-based company Alcor. If this technology arrives, however, it’s far from clear that it will be available to everybody. Many of the very wealthiest would presumably use it, but who else? Even if the technology got cheaper, where would the line be drawn? Would the severely brain-damaged be uploaded? Would we upload every gorilla? Every ant? Every plant? Every bacterium? Would the future civilization act like obsessive-compulsive hoarders and try to upload everything, or merely a few interesting examples of each species in the spirit of Noah’s Ark? Perhaps only a few representative examples of each type of human? To the vastly more intelligent entities that would exist at that time, an uploaded human may seem about as interesting as a simulated mouse or snail would seem to us. Although we currently have the technical capability to reanimate old spreadsheet programs from the 1980s in a DOS emulator, most of us don’t find this interesting enough to actually do it.

Many people may dislike this libertarian-utopia scenario because it allows preventable suffering. Since the only sacred principle is property rights, nothing prevents the sort of suffering that abounds in today’s world from continuing in the human and mixed zones. While some people thrive, others may end up living in squalor and indentured servitude, or suffer from violence, fear, repression or depression. For example, Marshall Brain’s 2003 novel Manna describes how AI progress in a libertarian economic system makes most Americans unemployable and condemned to live out the rest of their lives in drab and dreary robot-operated social-welfare housing projects. Much like farm animals, they’re kept fed, healthy and safe in cramped conditions where the rich never need to see them. Birth control medication in the water ensures that they don’t have children, so most of the population gets phased out to leave the remaining rich with larger shares of the robot-produced wealth.