Artificial intelligence, automated industry and colonizing Mars

From Marspedia
Jump to: navigation, search
This article is a source of controversy.

You can help Marspedia by discussing how this topic may be resolved.

Defining artificial intelligence

The sort of artificial intelligence[1] relevant to starting a colony on Mars is not the same type as human intelligence, such as the computer, HAL 9000, in the movie 2001: A Space Odyssey.[2] What is needed is the kind of computer program that runs numerical control machining[3] and industrial processes. The ability to mine ice and minerals and haul them back to the factory would include techniques developed for computer-driven cars.[4] Such an artificial intelligence would not need to talk with natural language but would be more like the expert systems doctors use to help them diagnose diseases. Talking would not be part of the computer's job and would be undesirable. An artificial intelligence should be expected to do specific things that it is programmed to do better than a human being, as the chess playing computer, Deep Blue,[5] did. An artificial intelligence may struggle at some things for which human judgment is beneficial. A chess playing computer has no feeling of joy for victory or sorrow for defeat. It is just as unconcerned if programmed to play poorly and loose most games as it is if programmed to win as much as possible. This lack of feelings makes a chess playing computer useful as a learning tool with a different roll than a human opponent. According to the introduction of Advanced Automation for Space Missions[6] machine intelligence has been used prove theorems, assemble machines with robotic manipulators, solve differential equations and write computer programs as well as the previously mentioned tasks.

A lesson from ants

There is good reason to believe that people can design a partially self replicating industrial system for Mars along with the artificial intelligence to run it following instructions it receives from Earth after every eight hour shift of Earth bound controllers. There are ants about three hundredths of an inch from end to end that none the less pack sufficient data processing power in their tiny heads to forage all about a human kitchen, find the cookie jar, suck the grease and sugar out of the chocolate chip cookies and return to their nest in wavering curving lines stretching over more than ten feet. They find a supply of water somewhere too, probably by digging into the dirt under the house's foundation. The task for an artificial intelligence running a partially self-replicating industry on Mars might be more complicated in raw processing power, perhaps ten times, perhaps a hundred times more complicated. People have powerful enough computers sitting on their desk tops. The things that are lacking are the programming and the design of the industrial task. These two must be built together as nature built the programming and the task of the ant together. People have an advantage over nature. We can see the final goal and design for that without going through all of the evolutionary steps along the way.

The challenge of the future

Colonizing Mars brings the task of advanced artificial intelligence controlled automated production to humanity while people still have a chance to successfully meet the economic challenge that goes along with it. If we put this off until the far future, it might kill us. We do not need this kind of advance in automation on Earth today because the economic problem is to distribute the fruits of production to those involved in the production. More people involved in production simplifies this task. The wealth we produce must be kept flowing like the sap in a tree, or the tree dies. That is a consequence of our current industrial economy which we must live with. The potential of automated manufacturing for producing goods without much input of labor could be too big of a change for Earth to adapt to. In outer space this advance could be helpful and might be necessary. When the technology gradually seeps back to Earth causing unemployment, the space habitats could be ready and waiting for emigrants from Earth looking for new homes. Space habitats could be flipping out of their factories like flap jacks, loading up with people and sailing off into the sunshine. The expanding economy resulting from people emigrating to outer space could soak up the increased productive power which otherwise might constipate the whole system. Mars is an ideal place for this to begin. People have written arguments like: " No significant fraction of Earth’s population will ever physically move off this planet." as if it were a statement from God and we must believe. In truth, there are physically possible means of transporting billions of people off of Earth.[7] The doubt is about whether or not people will have the political will and economic strength to make use of some of these possibilities.

Alternative industrial development

Why not send people to run industry rather than an artificial intelligence? People are very useful for controlling industry. Their intelligence is already demonstrated and does not need to be artificially programmed. However, people eat food, require drinking water, air and a host of special accommodations that are expensive to provide on Mars. People should run the industry on Mars but only after it has first been established by automatic machinery under control of an artificial intelligence directed from Earth. When the infrastructure is developed to support people, that is the time for people to arrive on Mars. Since industrial production on Mars will provide the recycling of air and water for people which is provided by nature on Earth, the ratio of industrial production per person will need to be higher than it is today on Earth. If people do not develop artificial intelligence as required to run an industry on Mars with only executive level instructions from Earth, the alternative of having people on Mars from day one of the colony is an expense to deal with in some other way.

Why not develop the Earth's moon which is closer and does not need this artificial intelligence? Luna could be developed also, but it has some severe shortages of materials needed for industrial production. Luna would need to recycle some materials with great efficiency and continue to import materials, perhaps from Mars. Developing industry first on Mars would make lunar development easier.

The options we have

Technological progress in industry is an unavoidable consequence of human nature. It is as though we were riding along on a train. We can direct the train to arrive safely at its destination, or we can ignore the train and end up in a wreck. Before 1700 the majority of people of each country were farmers. By the year 1900 Europe and the United States had made considerable progress in urban living and manufacturing technology. At that time, if we had simply forgotten how to use steam power and all the factories had stopped, people could have all gone back to farming without many starving because of the transition. If the industrial economy were to stop today, billions could die from famine. We can choose from possible futures. Refusing to enter the future is not an option.

There are not many attractive solutions to this problem. Outer space based manufacturing, with much larger access to materials and energy than Earth, provides an attractive solution to aim at. Artificial intelligence is one of the areas of potential advance that could lead to the higher productivity that might be necessary to establish colonies off of Earth.