Artificial Intelligence Capable of Generating Its Own Agenda:

The Point of No Return

Since Homo Sapiens emerged until recently, our life with other humans among nature, plants, animals, man-made objects and machines may not have been as peaceful and harmonious as we might have desired. There have been devastating natural calamities, big technological accidents and human conflicts escalating into battles and wars tragic enough to make William James summarize: “History is a bloodbath”.

But during all that hundreds of thousands of years there have always been a certain peace of mind: because we knew fairly well all the players that were of some significance for the events influencing our life.

Even if everything was changing all the time, through our evolution we grew up together with our natural environment, including plants, animals and climate. We know it is good to eat apples, and we know it is dangerous to swim among sharks.

We know our fellow human beings – individuals, or groups of individuals – like villages, companies, political parties and countries. And even if in our history there were innumerous examples of human behavior that was distinctly deviated from what was generally considered normal and statistically prevailing, with the time we have got used to that diversity too.

We, humans, are clever and we make our life easier by using tools.

neander

The Neanderthal man started with a club which we improved by designing more complex and highly sophisticated machines that extend our natural capabilities: in 21 century we move faster by cars, our doctor use CT scans to more accurately diagnose our condition, and the smart phone lets us talk to our beloved no matter where on the globe they now are, to mention just one of its hundreds applications.

In these circumstances the essential component in the complex peace of mind we still enjoy is based on only one thing: we can, and with very high certainty, predict events and their consequences for our life. Tomorrow morning the Sun will rise, the spring blossoms in the orchard will turn into fruits by the time when autumn comes, and when we put our card in the bank machine and dial the pin, we get the amount of cash we wanted. We do not get frustrated basically because we know the agenda of the things surrounding us, we can easily foresee what will happen and we know what pushing a specific button brings.  We feel we are in control.

One type of the machines we have designed perform a clear task WHAT to do by rigidly predetermined system HOW to do it. A simple vacuum cleaner could be an example.  A second, more complex level of machines are designed to perform a clear task WHAT to do, but can decide by themselves, following the criteria we defined, HOW to do it in the best way. Like our car satellite navigation, for instance. And recently the technology we call “artificial intelligence” accomplished an astonishing leap to a third level – the DeepLearning process where still we tell the machine WHAT to do, but through learning previous successful patterns and acquiring by itself new experience, the machine is capable of improving its performance indefinitely.  And, as it turned out, in the March 2016 five-game Go match when AlfaGo beat world champion Lee Sedol 4 to 1, the machine quickly learned to find better methods to win not through brute computer force, but by neglecting millions of bad moves. We observed for the first time in history a process in a machine neural network that is similar to human intuition.

alfago

The news about AlfaGo beating Lee Sedol sent shock tsunami to the humans all over the planet. But this was more of an emotional response, because the image we cherished about the game Go as a very special game reserved for human brain only, was a delusion. No game, provided it is played by clear rules, should be considered exceptional.

In this aspect, AlfaGo algorithmic system is nothing special. It is just one more extension of a specific capability. The Google team beat through its AlfaGo machine the World Go Champion in the very same way in which a person on a bicycle moves faster than a pedestrian.

And here is the main point: In reality, AlfaGo now is a champion. But it does not know it is a champion, it never wanted to be a champion, and it even does not know what the word champion means.  Because this level of Artificial Intelligence machine does not have its own agenda. It was simply instructed to play good Go, and it did so in a brilliant way of obedience. It is still we the humans who are telling the computers WHAT to do. So, up to here, no worry: the subordination dilemma does not exist – we, the humans, are still the indisputable Boss!

For now. With the current intensive dynamics in the sector, higher levels of AI systems are knocking at the door of our civilization every day, and their future functions could make things very different. Different to the extent to change the rules of the game, and to re-position the roles of the players.

There is a moment of time coming, and that moment of time will be with us probably sooner than we now expect: this will be exactly when a machine becomes capable of generating its own agenda – its own objectives to follow, parallel to or fully independent on any task, given to the machine by a human being.

The problem is that if an AI machine generates its own agenda, it will be an agenda that will be totally inconceivable to us. The texture of machine dreams and machine plans will be totally different from what we know about dreaming and planning. Which means their agenda will be for us unreadable and therefore unpredictable. For the first time in the history of humankind we will experience creatures with unpredictable behavior as part of our immediate surroundings – under the tree in the garden, next to the car, “playing” with the dog, checking something on the internet, taking care of the temperature for the kids’ bedroom or silently communicating with numerous electronic devices or other AI machines.

It doesn’t have to be always negative – maybe in the morning on your birthday you will open your eyes and, still half asleep, you will find on the bedside table your most favorite piece of art: Henry Moore’s “Pointed Torso”:

torso

AI knew you like it, found its image in the net, modeled it mathematically, arranged somewhere the 3D print and ordered a drone to deliver it. All that without you being aware of what has been going on for the last few days.

But the peace of mind that we have known for thousands of years, will be gone. More importantly, this would be the point of no return for all that has been defining the basic values that guided the advance of our human civilization for all its previous history.

Because one day then, when you open your eyes in the morning, still half asleep you may spot on your bedside table something very different from your favorite Henry Moore’s  “Pointed Torso”…

Advertisements

About lubomir todorov

The years of work and studies I have spent in countries of diverse anthropological, political and cultural realities like Japan, Australia, Russia, Netherlands, New Zealand, Czechia, and my numerous visits across the globe, were abundant in meetings and conversations with not only politicians, government officials and businessmen, but also people of different existential background, ethnicity, education and profession, philosophy and religion, social status and political inclination: aboriginal painters in Australia, US generals, Japanese Buddhist and Shinto priests, Majesties and members of Royal families, Russian scientists, street Mapuche musicians in Chile, British parliamentarians, Indian philosophers, Czech university professors, CEOs of top Japanese corporations, Chinese social sciences researchers, Dutch entrepreneurs - to mention just a few. Over time all this primary data, intertwined with the everyday inflow of information about human activities all over the world and their real-time consequences, were re-assimilated in my mind through the tools of philosophy, international relations, ethics, biology, economics, history and political sciences to gradually constitute a distinct resolution to re-examine my understanding about politics and the nature of human society. Making all the time hard efforts to keep the methodology machine uncontaminated by ideological, political or personal prejudice, I clung to one rule: If you aspire for the beauty of truth, it is only facts and logic that matter. Questioning the viability of human social and political behaviour and the capacity of existing political systems to lead to where people want to be, I believe that sooner than later, in pursue of sustainable prosperity, our humanity will embrace its ultimate imperative of Civilizational Thinking. Human Civilization is the spiritual dimension of homo sapiens group survival strategy that urges humans to mutually defend their strategic self-interest by generating Civilizational Values; and the optimal political environment for that is Universal Future: a global multifaceted platform on which all polities in their ideological, political, national, cultural, ethnic, religious, racial, etc. diversity exist, interrelate and compete with each other on non-violence basis.
This entry was posted in Artificial Intelligence, Brutal Logic, Humans and tagged , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s