Listen to the podcast here.
Artificial intelligence (AI) is a hot topic these days. Some predict humanity will see superhuman AI by 2020, and people like Elon Musk view it as a major threat to watch out for. Earlier we wrote about AI because most people don’t know what to look for, so most people can’t see what we have achieved already. I will repeat part of the earlier post here.
First to understand AI, one needs to get a grasp what intelligence is. To me it is ARGO, Autonomous Robust Goal Orientation. Intelligence is a property of a class of systems, including humans and machines. The system classes are
1. The Tool (can be anything)
2. The Cybernetic system (a mechanism that has an internal target parameter and responds to deviations from that parameter)
3. The Intelligent system (A system that strives for a goal, which is a specific combination of parameters, The more autonomous and robust the system is the more intelligent)
4 The Aware system. (A system that can adapt and prioritize its goals to fulfill its fundamental needs)
An electric drill is a tool, if it has systems to maintain RPM and manage battery life it is a cybernetic system, if it does so robustly it becomes borderline intelligent, and if it would selectively distribute perception to detect wheter it should optimize RPM or Battery life at any time it would be rudimentary aware. In most cases though an electric drill lacks robustnes to be intelligent, so awareness would be very fragile.
Robustness is perhaps the most significant aspect that is commonly ignored in AI research. It matters not whether a computer can solve a complex problem, or talk like a human, althought there is a relation between a choice of viable responses and the number of independent ways a response can be produced (robustness). The human brain is vastly redundant. It is build out of neurons that can die for many reasons, and are expected to do so without risk to the person involved. Open up a computer and stick a screwdriver in the wrong place, and the whole thing grinds to a halt. The same goes when a line of code running AI contains an error. The same even goes when a neural network is implemented on a massively parallel computer. We would be seriously scared if a computer manages to survive and achieve its purpose even if we try hard to stop it, that is not how computers behave. Robustness is what super human AI interested are so worried about. apart from the fact that the goals set by the AI most likely don’t align with the ones keeping us going. The threat of an autonomous robust AI with any real power to manipulate our world is real, just think of the animal kingdom as examples of ARGO. How many animals are friendly to us? They are either scared of us or ready to eat us.
But the people enthousiastic about the prospect of AI often ignore the AI we already have, that we basicaly inhabit. Because most people don’t view humans as mindless drones or computational resources (except perhaps Amazon which launched a mechanica turk service). Humans are aware organisms that can adapt their goals in order to make them align with their fundamental needs, our motivation to drink, eat, have sex etc. It is probably true that many animals can be demonstrated to have the ability to strive for a goal that has nothing to do with their normal behaviour in their natural habitat (even though they would be limited to their motoric repertoire). Humans can be trained to strive for goals that are symbolic, that don’t satisfy any need right now. Therein lies a weakness in the integrity of the human mind, because for those that know how to condition responses to symbols of future ‘need satisfaction’ the possibilities are endless. Examples of symbols most people are conditioned to are banknotes and coins. We can’t eat them, they don’t give us shelter, the don’t give us satisfaction, but they symbolize our power to get all those things. This is the problem : Most of us respond strongly to a symbol that privided it is promissed or supplied in the right quantit can make us accept almost any goal.
It is enough of a threat to our sanity that we can be conditioned to strive for symbolic goals. This already makes us a material for AI without awareness, in the role of mindless drones. Awareness is a process of information amplification that allow us to distinguish between behavioural options (relative to our knowledge and needs) or goals. If we accept goals that we don’t generate ourselves, we can turn into servants for an awareness that may not take human survival into account. Domination of slaves by slavedrivers is a good example. The master wants the cotton, the slave may die in the field.The masters mind being as warped by symbolic greed as the slave’s by elimination of free awareness.
Language and words are symbols, but that does not mean words and language are a high risk, because what determines the risk of symbols is how we are conditioned to respond to them, If someone you believe comes up to you and tells you you can become a milljonair if you stand on your head, it is you being conditioned to the implications of being a miljonair that will (probably) make you stand on your head. Our leaders are people that have conditioned us to respond to their their symbolism, usually by adopting language we have been conditioned to, other times by using symbols of intimidation, fear and danger to give us the proper conditioning response. Conditioned words are a currency all by themselves.
The power of other symbols like money hands a similar conditioning tool to the people that control it. It is in their interest to condition us to this symbol, to make money essential in our lives, to make it important in our minds. This is done on the one hand by putting a price on everything we may be interested in, and making it hard to come by. On the other hand we are conditioned by the marketing industry, selling us lifestyle goals, materialistic ideals, social pressuring us to own tons of useless stuff, all the time using our fundamental needs such as the need to belong, our need for freedom.
When the question is asked “Does our economy constitute an AI” then the answer must be yes. This is easy to see, because our economy has goals, namely the growth of GDP, and many secondary goals set by different organizations that are part of its system (mainly the financial and banking system) have nothing to do with our needs. Within the conditioned symbolic framework of ‘the economy’ we are expected to cheer when banks do well, when Wallstreet is up, and mourn when export is down etc. While some causal relationship between the wealth we thus help to create and the part of it we recieve exists. No relationship between our health and security and the achievemet of symbols like a greater GDP are guaranteed, in fact, we see countries with decades of ‘growth’ sink into sudden dispair. Economic goals are alien to us even though close to us they overlap with basic barer and trade, sharing of effort in production of wealth. But economics is not about that we will see.
When the question is asked “Is our economy aware” the answer is also yes. It is adapive to its substrate, which is sufficient fossil fuels supplied to the manufacturers, ensuring cashflow and profit which ultimately determines the GDP and other values of economic theory. It is aware by borrowing human awareness, moments of abstract thought. An analyst with a bank will analyze a specific region for possible profit from logging for instance, driven by many conditioned goals that have at their pinacle some economic prosperity ideal for locals, but also the analyst salary. The more abstract, materialistic and egocentric the analyst, the better he/she will function on the job, because the easier it is to utilize his/her mind.
Awareness is te ability to adapt behaviour to a chosen sensory experience out of the total at any time, be it through speech (saying ‘I see a bird”) or some other motoric activity (like pointing to it). We are aware of something before we commit to an action, then once we act and we become aware of new things amongst which the experience of the act. What awareness does is modulate our perception in a dynamic interaction with all our other (relevant) goals and constraints so that a decision about what action tot take results. Once we have processed a situation with a certain resulting decision a number of times we lose awareness of that situation when it occurs, our behaviour becomes automatic. The AI called economics is aware because it does analyse, through the use of analysts, computers and economists, what the best option or goal should be in order to sustain its dominance, and if the data is insufficient to make a decision more data is sought and different opinions and options are weighted against each other. When the human (and increasingly silicon) symbol manipulators finally come to a clear advantages of one of the actions, they become policy, in the interestof the survival of economics.
Sure there are many individual beneficiaries of economic theory, but they are interchangeable, the bank CEOs, the economicsts, directors etc. all take away a stash of ‘symbols’ which in itself conditions others to want to do the same. ‘Stealing’ bankers make us angry because WE want that money. While really we don’t want to need it, which is something different. But economic theory and economics is not controlled by anyone, that makes it so robust, It is a system to be trusted and participated in, one religiously defended by people that don’t even benefit from it. It is an aware Super human AI that controls large parts of humanity.
It is not principally necessary for people to disengage from this AI as long as the overal trajectory it sets humanity on is aligned with our fundamental needs, so at the end of the rainbow there should be food to eat, water to drink, grass to lay down on, and life to enjoy. Sadly this is not the case. We are stuck with a system that is achieving the opposite, even if many of us are fine at the moment. We are lending our cognitive abilities to a mechanism that does NOT care about us because it only looks at abstract parameters, because the humans that do its bidding have been conditioned to want money, no qualitatively superior outcomes for themselves and others.
Economics has an Achilles heel though, namely the requirement that plenty of fossil fuel is available. A short definition of economics is “A theory to maximize the utilization of fossil fuels” which was considered a harmless goal to bring prosperity to the masses using a plentifull gift from the gods. No wonder economic forces work against renewables. Without fossil fuels there can be no sudden increases in productivity and manufacturing capacity, no cashflow, no credit. The unique property of a fossil fuel is that it is dormant production power. It is storage in itself. Like having a massive charged clone army that can be let loose on any problem or ambition at any moment. Production capacity and speed can reach any value as long as there is enough fossil fuel credit. Hence in our fossil fuel based economy one can suddenly deploy many workers and machines because the energy is waiting, stored, ready to be activated with ‘carboncredit’.
Right now it is clear that economics and free market theory (which is not praticed most of the time) is not leading us to a more prosperous and secure future. The main culprit is the abundance of fossil fuels, which removes limits on the speed at which resources are harvested. The only limits are 1. fossil fuel security and international distribution of its consumption and 2. the financial system, so the availbility of credit to distribute the fuels. But the world economy is like a lawnmower that is programmed to never stop, and you and I are its computational minions.
Because economic goals are often aligned with individual goals of the people involved (when they earn enough or believe to have the chance to do so) it is hard to deprogram enough people to return to a more sane goal set, one produced by human values, not money symbolism. Many people believe they can do social and climate ‘good’ using economic principles, while in fact, they can’t because they (who may not be a serving of the Economic AI) by lending and making profit and importing and marketing etc., feed into the AI that has thousands of people working for it, maximizing cashflow, utlization of fossil fuels, depleting resources where they can be found and moving on like a buch of locusts.
The answer to resouce destroying economics is extraeconomics
To turn this around the rules of economics need to change, so that people that are captured by its incentive structure, that will work to give its goals and ideals life, do not in fact clearcut our planet and ruin prospects for humanity. It is impossible to fight the AI because there is an overlap with the real economy that hide the destructiveness of the marcoeconomic effects. It has become a part of many minds much like some medieval bloodletting in its day, precieved to be a healthy burden.
Rule change : Maximize life, not cashflow or GDP. This means all life. Resources can be used, and value of them be accounted in the usual economic way, but the value of a project or asset depends on its ability to increase the variety and abundance of life (lets say natural to the region), and credit goes to those projects or companies that maximize life the best. Life of people, animals, plants, oceans..
Maximizing life will also minimize destructive practices like the use of fossil fuels, and promote energy harvesting methods compatible with life like wind, solar, tidal etc. It calls for maximizing the amount as well as the variety of species, something that will disadvantage fossil fuel intensive industrial farming and prepare the species of earth for the selective process of global warming.
The choice when allocating resources to one project or another, should be based on the ability of each to increase the variety and abundance of life.
The Life Maximizing Economy could have a currency representing life, and have lie futures that can fund reforestation, ocean restoration and other projects (meaning take a piece of the fossil fuel pie to start them up). Of course the threat of old Economics AI looms large as long as fossil fuels are used, so banning and phasing those out should be priority nr. 1. Once selling as much fuel as long as possible is nobodies goal anymore, production will start to depend on humanities fundamental needs again.
Maximizing life is not an incongruous goal, after all, economics always lies that it tries to maximize ‘growth’. It thus claims to do wat we should make it do.
It is simple, we may be hijjacked to serve an economic philosophy that is killing off life on this planet, but we are also capable of adapting the program in our minds to make it secure for us, our decendants, and all other living creatures on Earth. Thus we will learn for the first time to recognize and defeat a Aware Superhuman AI.