Thursday, April 08, 2010

The Dawn of the Intelligent Planet

IBM Forum Slovenia 7-8 April, 2010
Keynote Speech by Hardy F. Schloer of Schloer Consulting Group





The Dawn of the Intelligent Planet


Ladies and Gentlemen, good morning

First of all, I’d like to thank the organizers of this event, for inviting me to speak to you this morning.

Today, I’d like to share with you some fundamental observations that should lead to a better understanding of how the future of technology must, and will likely, unfold, so that our planet can evolve in a more positive way.

My talk today carries the title: The Dawn of the Intelligent Planet

We can draw from this title three implications:

1) Something important is about to begin
2) It involves the concept of Intelligence, and
3) It doesn’t concern only the future of our local world, but also that of our entire planet!

Today I’d like to talk to you about the next important evolutionary step of humankind, which, I believe, will be known in the history books of the future as the period of transition from “Human Intelligence” to “Human-Machine Intelligence”!

We are all here to participate in this amazing experience that will hopefully change the future of humankind in a radically new, positive way. Now, let me make a rather provoking statement:

The only way of ensuring a safe and healthy evolution for our planet is by handing over all of our vital decisions and planning tasks to intelligent machines. Full Stop!

Now, some of you may immediately begin envisioning large packs of dangerously out-of-control, “gone crazy” robots that will roam the planet, enslave the world, and end the freedom of humankind as we know it. As a former clinical psychologist, however, I can tell you that if these fearful thoughts do cross your mind, you are clearly “projecting.” That is, you’re projecting your own predatory human nature onto machines…

But let me allay your fears at this point: No machine, no matter how intelligent, will ever be as unpredictable and dangerous as humans are today. Computer intelligence, as it will develop in this new era of human-machine intelligence, will be much safer, reasonable, and predictable than we, humans have ever been, at any time in our history. Just consider this: You can take a hammer, and smash a notebook computer in front of 50 other computers, and they will not even care, much less attack you for it…

If you are not convinced yet, consider, for a brief moment, the collective human behavior of the past few millennia! I believe the record speaks for itself: Humans still are as they have always been: lethal predators, eager to kill, tireless seekers of opportunities to expand their power and possessions, regardless of the endless misery they inflict on themselves and others.

Man has not changed! Not since the very early dark ages of humanity, some millions of years ago, when it all began. He is still the most dangerous and rapacious of all the creatures on the planet. Along these lines, man has built social, political and economic systems, which he has named ‘civilization’ and ‘civilized behavior’, but which are nothing more than very clever ways of practicing his ancient instincts of hunting and killing. To give a modern example, ‘man’ has invented financial markets: a zero-sum game, where one can gain only at the expense of another, with no regard to the collective detriment or the high cost that one’s gains may inflict on everyone else.

And yet, man also has the most amazing and generous creative abilities, having produced, over the centuries, unequaled artistic and technological masterpieces that have indeed the potential of changing, for the better, his life and that of all the other creatures on this planet. Man has also invented machines. But like men, these machines can do nearly everything and more. They can build or destroy; communicate or hide in opaque secrecy; they can calculate, predict, make a space vehicle land on Mars with the precision of a square-meter, or deliver a deadly missile across 1000 miles into the bedroom window of an enemy…

Now, the question we must ask ourselves is this: with more than six billion humans on the planet, equipped with these powerful technologies that can destroy or built up the Earth’s valuable infrastructures, how do we manage the ever accelerating evolution of more and more effective machines, given man’s unquenchable thirst for domination?
How do we manage man’s willingness to engage in conflict, even if it takes his own life, or that of millions of others? Can we trust man in this more and more complex world to make local decisions that have global effects? Human management of this planet is truly in question!

Something is really wrong here, because the human condition is not improving, in spite of all our wonderful abilities and beautiful innovations. On a global level, hunger and poverty are vastly increasing; economic distributions on a global scale are dangerously unequal, and even in the best of societies we have lost the sense of what is truly valuable in life. We are increasingly the slaves of communication devices, overwhelming information systems, and technology structures that have not adapted to human needs but, rather, forced humans to adopt their lifestyle to the intrusive technological infrastructures of this planet.

Only a few months ago, I attended a public debate on the question if it were possible to live a week without any communication devices: for one week simply go back to the life we lived in the 19th century, with no phone, no TV, no Internet; spending time with the family in the evening, sitting together around the dinner table, and perhaps, reading from an interesting book to the family, by way of collective evening entertainment.

During this debate, a lady in her late 80’s, a former professor of history, explained in great detail how it all was in her childhood. Things happened very slowly then. People still had time to think. People lived in some ways a harder life, but they were happier. The mental illness of depression was much rarer in those days. By the end of the debate, most agreed that a perfect vacation would be one without any technology… going back to basics… being “off”, rather then being “on” all the time.

For me as a dedicated futurist and technologist, this was a clear sign that we are still missing a very essential ingredient, before technology becomes a true and positive catalyst in human development. In other words, we need to develop the next step in the evolution of humankind, the Human-Machine Intelligence. What is at stake here is no less than a process of co-evolution, in which humans and machines will become partners in creating a new mentality and better forms of life for everyone concerned.

Let us remember that the development of technology is in fact a leading component of global human development. We MUST become aware of and live up to the exigencies of this new form of symbiosis and co-evolution. In this room, we all are closely related to technological development and its innovation. We are the people that are accountable for the very important transitions of humanity in the future. We must live up to this responsibility, today and in the future!

So let me summarize my argument up to this point: if we are to usher in the dawn of Planetary Intelligence, we need to attend to the co-evolution of humans and machines. I’ll now look, first, at the machine part of the equation.

One of the fundamental problems of today’s technology is that it still requires the participation of humans to function. In fact, we are occupied more than ever to interact with all of these devices and machines that we have build, rather than let them just do their work automatically. Technology is absorbing us, rather then helping us.

Machines can do things very fast, but humans are very slow in evaluating the results of their performance. Again, the financial industry is a good example here, where computer-based, algorithmic trading becomes more and more a decision-processor in the millisecond space, where billions of dollars are being exposed to transactions that can no longer be followed by humans in real time.

On the one hand, machines have evolved to a point where they can do substantial tasks, and act and react at speed levels that are highly uncomfortable or even intractable for humans. On the other hand, the machines of today still lack true intelligence, and therefore, they can cause substantial disasters, usually at high speeds that can substantially magnify the negative results.

But we must also recognize that computing machines are beginning to close the gap between learning and acting. To refer, again, to the financial industry, computers now begin to read all global news automatically, analyze its content in milliseconds, and deploy trades instantaneously. But, of course, there is a disastrously weak link in this technology: the logic of executing trades is hard-coded by the so-called Quantitative Analysts. Herein lies the whole problem: should there be a change in the nature of data input, then the interpretation of these news feeds must change as well. But the current generation of machines can’t react, they are not allowed to deviate from their hard-coded instructions! Therefore, these machines are only half-smart. And that’s the point!

What is needed is intelligent machines-- not hard-coded rules, but true machine intelligence.

Current artificial-intelligence systems are, as a rule, one of two types: logic-based or probability-based. But researchers, including myself, have developed lately new technologies and computer languages such as MIT’s Church, or my own Quantum Relation Technology Language called for short QRT, that combine the best aspects of each type, and make AI smarter, more humanlike.

It started with AI researchers, back in the 1950s, who thought of the human mind as a set of rules to be programmed. Thus, they developed systems based on logical inferences: "if you know that birds can fly and are told that the Eagle is a bird, you can infer that Eagles can fly."

But with rule-based AI, every exception had to be accounted for. And we learned the systems couldn't figure out that there were types of birds that couldn't fly; they had to be told so explicitly, by coding these exceptions into the program. Later AI models gave up these extensive rule sets and turned to probabilities: "a computer is fed lots of examples of something - like pictures of birds - and is left to infer on its own what those examples have in common."

Church and QRT are both "grand unified theories of AI" with both systems creating probability-based rules that are constantly revised as the system encounters new situations. For example, a Church or QRT program that has never encountered a flightless bird might, initially, set the probability that any bird can fly at 99 percent. But as it learns more about the Ostrich or the Penguins, and caged and broken-winged birds - it revises its probabilities continuously. Eventually, the probabilities represent most of the conceptual distinctions that early AI researchers would have had to code by hand. But the system learns those distinctions itself over time - much the way humans learn new concepts and revise old ones. In the early years of my research in defining new models of AI, I also called this approach “Human-Emulated Artificial Intelligence.”

Today we know that these new approaches surpass already current AI models. Newly developed applications in which, for example, a QRT system was deployed to make predictions based on a set of observations, did a "significantly better job of modeling intelligent returns, than traditional artificial-intelligence algorithms did in the past.”

Of course, these new technologies still need further improvements and specific operations are extremely "computationally intensive" when they tackle broader-based problems. I am sure that the Hardware division of IBM is delighted about this fact, because it insures the prolific sales of supercomputers well into the future.

Nevertheless, this is only the beginning!

New systems must begin to model global problems, and must have the ability to understand and process interdisciplinary problems in parallel, internally and continuously. Such global systems must have the ability to contain all local problems within; they must be globally connected and must fully account for the “butterfly effect.”

And this, ladies and gentlemen, requires the building of a fully interconnected and intelligent planet! This, of course, also brings me back to the beginning of my talk and to the other term in our co-evolutionary equation: the human factor.

I’d like to say this once again:

Intelligent machines should not only solve systematic local problems, more importantly, they must become an important and responsible part of global human development. As one of my close friends and collaborators, Prof. Mihai I. Spariosu from the University of Georgia, in the US, has argued, global intelligence is the ability to understand, respond to, and work toward what will benefit all human beings and will support and enrich all life on this planet. Global intelligence is based on the collective awareness of the interdependence of all localities within a global frame of reference and the enhanced individual responsibilities that result from this inter-dependence.

As no national or supra-national authority can predefine or predetermine it, global intelligence involves long-term, collective learning processes and can emerge only from continuing intercultural connectivity, open dialogue, and peaceful cooperation of all members of the planet.

The phrase “what will benefit all human beings” in this context, however, should not be understood in the utilitarian, restricted sense that implies the excessive, materialistic focus over the wellbeing of humanity in general. The new models of global intelligence will sooner or later give humans back their freedom to no longer be overly concerned with the management, or even the productivities of this planet, but with the responsible enhancement, stewardship, and enjoyment of its beautiful gifts.

The science fictions of the 60s and the 70s in the last century always envisioned the year 2000 as a futuristic society where computers did all the management, and the robots did all the productivity. But the year 2000 came and went… and there is still no trace of such a society. What is still missing is this “unified theory of intelligence” that would enable us to build our societies, based on global intelligence, which is in turn based on the co-evolution and symbiosis between man and machine.

We are now working towards such goals. But we must not stop developing the tools needed to get there.

So, let me highlight some of the basic technologies and infrastructures that need to be developed to bring about the “Intelligent Planet” in the foreseeable future:

(1) We need to develop massive, supercomputer-driven, global knowledge centers that manage all of the Earth’s open-source data globally, and analyze its content in an interdisciplinary and intercultural form.
(2) We must connect all these global knowledge centers, so that they can become a globally connected mash of knowledge depositories
(3) We must also develop a global mash of networked sensory devices and data extraction technologies that collect information of any kind 24 hours a day, 7 days a week and transport such information in real time into the knowledge depositories
(4) We must develop the best AI solutions possible, to continually search our knowledge depositories for deep-rooted patterns and understandings that point to globally responsible opportunities and common planetary risks
(5) We must build access technologies that allow anyone at any time to access these analytical knowledge depositories and use its information at no cost, to allow human development independent from economical power
(6) We must build broadcast technologies that use a world-standard format, and continuously broadcasting streams of data and information, which include any type of risk or corrective information that may be vital to human development
(7) We must develop these solutions as global utility that remains free of charge, and free of all political and dogmatic influences
(8) All vital human services, such as global commerce, healthcare, education or even governance must become a global solution with local subsets
(9) All local machine intelligence must have full access to these automated global knowledge networks
(10) All global information must remain open-source knowledge, available to all members of this planet

Implementing these 10 points will be the vital base for the Intelligent Planet.
And finally, we must realize this:

Today, there is only one serious technology player left on this planet that can take us to this future, that is, the future of the Intelligent Planet. The player I am talking about is IBM. It will take astronomical amounts of investment into newer and faster hardware technologies, comprehensive commitments to develop new global middleware and interconnecting mash technologies, as well as other similar systems, to work toward this Intelligent Planet

But one thing is clear: We must make the Intelligent Planet our most important goal. Given the complexity of modern society and the desideratum of continuous, peaceful human development we MUST work toward its successful accomplishment, and not be stuck in debates about its necessity.

So, I encourage everybody in this room, to be part of one of the most challenging, but also most rewarding frontiers of our millennium: The Dawn of the Intelligent Planet.

Thank You!