Machine intelligences are all around us in society, so why is no one talking about them? With many prominent figures in the world of computer science talking about how we may have Artificial General Intelligence in our world in 20-30 years is seems strange that no one is raising the bells of alarm. Well, actually they are. Reports are being put out by places like the Oxford Future of Humanity Institute, and the Machine Intelligence Research Institute, which not only go into the mathematics behind aligning our AI’s with our values in the case that become more intelligent than us, but also in the strategies that we should employ in our research towards this goal of AGI. If there is one word to describe how we should proceed, it is carefully. There are a myriad of uses for AGI not including controlling the entire means of production through machine intelligence and computation. This is the case we will examine today.
Imagine for instance that a machine was capable of controlling the entire economy. Yes, I do mean the entire economy. From production to distribution, filing transactions and marketing, to closing mergers where economically optimal to maintain efficiency. Sound unlikely? We may be closer than you think. This is what researchers have in mind when they mention terms like AGI. An AI that can do everything we can do, but better. So what role will humans have to play? Granted there will be a long transition period (see my work “On Artificial Intelligence” for a technical explanation of how work is changing now) but it seems that soon our world will be a lot different than it is now. The questions to ponder if you are questioning this reality would be: Would a machine be better at creating and operating new machines than us? Would a machine be able to route delivery trucks better than us? Would a machine be able to create content and entertainment that we would like? Think of the jobs that have already been automated and the amount of work today (virtually none) that is done in office buildings. In fact the only work that is done today is of two classes; manual labor which may take a long time to be overtaken by machines, and human resources oriented white collar jobs (most of the rest, including some aspects of law can be automated). Now I am not saying that there will not be a place for humans, I am just saying that the world will be a very different place in 20 years, and the nature of the changes may take us a long time for us to realize, as a society, what is going on.
Some issues pressing researchers now are; what to do if a small group of individuals makes an AGI and as a result gains skyfall profits? How will we ensure that they do not hold the rest of humanity beholden to their every whim. This brings a whole new meaning to the term “class struggle.” However as of right now AI is quite benign. Most intelligences use linear or quadratic, or even some use higher order approximations to the truth in a statistical language that is hardly representational of the way humans think. But the changes will come faster and faster as the nature of AI research is such that it builds on itself, accelerating and accelerating so that the machine will build and improve on itself. Ever expanding into new domains which were once thought the realm of only humans. What will happen to the human race? Join me as I embark on a research journey across space and time to catch a glimpse of the cutting edge in AI research and to discover the future of man on earth.