Chapter 12: Artificial Life from the book "Life Begins Here"

Today, researchers are trying to build new life in the laboratory, and look for, at least in theory, non-carbon-based life that might exist somewhere in the universe

Life as a concept

While the echoes of the classic debates about the feasibility or impossibility of artificial intelligence were carried in the air, the supporters of the artificial beings have already moved the fight to the depth of the area of ​​the debate about the essence of the phenomenon of life. The creation of new life may turn out to be possible - if you are willing to define the phenomenon of life as objectively as possible, without relying specifically on the carbon-based life that developed on Earth. The researchers involved in "creation science" are actually trying to find and formulate a definition for the phenomenon of life, which would include the carbon-based life within which humans evolved, along with a wider variety of phenomena (partly theoretical), which may exist in different regions of space-time. In other words, it is an attempt to see the phenomenon of life that we are familiar with as a particular case from a much broader set of possible phenomena (similar to the way in which the theory of relativity includes within it, under certain conditions, Newton's laws).

According to this concept, the organisms living on earth today (from humans to the last bacteria) are nothing but very complex biochemical machines (and perhaps a collection of many such complex machines and systems). As Adi Pros says: this great complexity gives the animal systems a "non-linear" behavior, that is, not self-evident and not easily calculable and predictable. This "non-linear" behavior is considered one of the most prominent characteristics of the phenomenon of life. The researchers trying to create living systems ask if it is possible to create a feature of "non-linear" behavior, not through very complex systems, but through many relatively simple systems (for example, computer programs), which manage a constant and very complex "talk and achieve" with each other.

Life, according to carbon-based biology, begins as mentioned with a molecule capable of organizing its environment and replicating itself. The existence of molecules with such properties leads, over time, to the formation of living cells that are later organized into tissues that form organs that make up the complete system, which is the animal - whose behavior, as mentioned, is not "wired". A system of artificial life may reach the same "bottom line" by a different route: simple machines operate according to simple "laws" or "recipes" and together create simple structures, the generalized dynamics between which determine the behavior of the overall system - which may be very complex.

This abstract-something description can be illustrated using the model called "Automation Thi", proposed by the mathematician John von Neumann (1903 - 1957 John von Neumann), who is known among other things as the builder of the first electronic computer. The model is based on a chessboard-like pattern, which has a very large number of squares. 29 "states" are defined on this board (such as "happy", "sad", "quick"). Transitional conditions are also defined. These conditions determine how the current state of the "artificial creature" (which is a collection of contiguous cells acting as one entity), in relation to its relative position on the "status board", will affect its state during the next "generation". For example, if your status is X, and you are on the board at position Y (in relation to other "status" squares), then in the next turn your status will be Z. Any change in one of these factors affects the state of the "organism" in the "next generation". Running this model on the computer showed that within its framework there is a possibility of the development of a "self-reproducing machine" that emits from it a structure identical to itself, which in turn emits another identical structure, and so on.

John Conway (1937 John Conway) from the University of Cambridge, England, developed a similar system known as the "Game of Life". This system is based on two states ("alive" and "dead") and four simple transition rules on a checkered board, and after running for a certain period of time, it enters "permanent flashing". These descriptions (and especially John von Neumann's cellular automaton) are reminiscent of "Feynman machines", named after the Jewish physicist, Nobel laureate in physics, Richard Feynman (1918 - 1988), who invented them. These machines, Feynman said, would be programmed to create small automatons, which would be programmed to create even smaller machines, until they evolved and created molecular-sized machines. The MD writer Arthur C. Clarke based his book "The Face of the Future" on such machines, capable of operating at the molecular level. The book describes a machine ("replicator") capable of taking any material and producing any other necessary material from it. This is how you can end hunger and poverty in the world (and you can also do some less positive things).

The first "replicator" will indeed cost a great deal of effort and a considerable amount of money, but the second and third will already be built free of charge by the first "replicator", which will reorganize its environment and replicate itself. In other words, according to the accepted biological definition itself, he would be considered an animal for all intents and purposes. The "Reflector", "Feynman Machines" and "Neuters" created by Theodore Sturgeon in the story "Al Bezair Anfin", can be seen as links in a chain that includes Carlo Collodi's (Lorenzini) Pinocchio, Mary Shelley's Frankenstein, the doll Olympia from "Hoffman Stories" ” and many other stories.

Computerized evolution

During the writing of this book (August 2015 - December 2017), we learned of the death of John Holland (1929 - 2015 John Holland), one of the pioneers of the field that later earned the nickname "genetic algorithms", or "computerized evolution", or "evolutionary programming" . Holland, who was a cluster expert, a professor of psychology, electrical engineering and computer science at the same time, must have been familiar with the pioneering computer work of Tibor Ganti on his "Chamoton" model. It seems that all these skills served him in a combination that seemed unnatural at the time, between technology and mathematics on the one hand, and between principles of evolution and life sciences on the other. The breadth of knowledge of the scientists who worked in this new field, in those years, the 60s and 70s of the 20th century, also stood out in the case of Lawrence Fogel (1928 – 2007 Lawrence Fogel). He received a doctorate in biotechnology for a work entitled "On the Principles of the Organization of Reason". He later worked in the development of rockets and also participated in the development of the "Atlas" rocket, which at the beginning of the race to the moon carried manned spacecraft into orbit around the Earth.

Holland and Vogel laid the foundations for the ability (which developed over time) to harness the process of genetic fusion, which has already proven its effectiveness in the living world, to the cart of artificial "creatures". With this technique, genetic algorithms (software recipes) have been written ever since, or in their broader name, evolutionary algorithms.

Such algorithms may offer ways to obtain solutions to problems for which ordinary computer systems require a great deal of time and computational volume to solve. Basically, evolutionary programming - like evolution in the animal and plant world - places its gold on luck, which should be followed by competition and selection. To apply this technique, one randomly selects (for example, by tossing a coin) several possible arrangements of the "switches" that represent information in the computerized system. The computer responds to each arrangement with a certain action, which can be expressed numerically. Thus, there are arrangements that yield a low numerical result, and there are those that yield a higher numerical result. Assuming that "the more is the better", we can say that nature tends to give advantages to arrangements that yield higher numerical results. With an uncomplicated calculation method, it is possible to create a population of arrangements, which will express this preference for the arrangements with the highest numerical results. That is, there will be more arrangements with a high numerical result, and fewer arrangements to which the computer responds with a relatively low numerical result.

At this stage, pairs of "successful" number arrangements are created (randomly), interrupted (again randomly) and organs, or "materials" are exchanged between them, in a way reminiscent of the interruption and integration of chromosomes in the reproduction process in biological systems. As strange as it may sound, repeating this process, with the arbitrary ("divine") intervention of the system operator and changing various signs in the structures of the "breeding" programs running on the computer, creates mutations and yields a series of increasingly fine (higher) numerical results. In this way, it is possible to solve complicated problems that are very difficult to solve in an analytical way. Among these we can name systems that try to simulate the complex behavior of living "creatures".

While Holland and Vogel were operating in the United States, Ingo Rechenberg (1934), at the Technical University of Berlin, developed a technique he called "evolutionary strategy". This computerized system was used by him, among other things, in improving the efficiency of wings in theoretical aircraft, as well as propellers, turbines and impellers for utilizing wind energy. Rashenberg was also known as a broad-minded person. He was interested in geology, archeology and zoology, brought rock samples from the desert in Morocco, and even had the Moroccan "flik flak" spider, which is characterized by surprising movement ability, be named after him: Cebrennus rechenbergi.

The forefront of contemporary research in this field passes through the Institute for the Study of Complexity and Complications in Santa Fe, of which Lawrence Vogel was a member of its board of directors, and whose founders also include the Jewish physicist, Nobel laureate in physics, Marie Gell-Mann (one of the developers of the standard model of the structure of matter). In the conferences held there, quite a few computerized systems based on simple initial structures, and on independent running, without operator intervention, were presented, which succeed in creating various characteristics of living biological systems. One system moves aerodynamic objects in a way very similar to the way birds behave and fly in a flock. Adding beeping sounds to the image projected by the system on the computer background may create an almost perfect illusion of a flock of live birds. This system is defined by only three basic laws. Each of the "birds" moves in space completely independently, and yet, between all the "artificial birds" a dynamic of a flock of flying animals develops over time.

core wars

In the 90s of the 20th century, Thomas Ray (1954), at Michigan State University, USA, developed a computerized system that simulates the development of artificial life. He called it Tierra, and within it computer programs competed with each other - just like living creatures - for resources. As in an earlier system, called the "core wars", also in Tierra the resources that the "software creatures" fought over were the central processing unit (CPU) operating time and access to the shared memory pool. The competition for the scarce resources caused a kind of evolution, when natural errors led to the creation of a kind of mutations, some of which gave their "owners" advantages and some - disadvantages.

About a decade later, in the early 2000s, and as a result of Tierra's inspiration, Charles Ofria (1973 Charles Ofria) and Chris Adami (1963 Chris Adami), at Michigan State University, USA, developed a more advanced system in this field, which Her name was Avida. This system - which was based on an older system they developed together with Titus Braun (1974 Titus Braown) at the California Institute of Technology Caltech - has already been defined as a "software platform of artificial life", and the "creatures" created within it are capable of replicating themselves.

In other systems, a kind of simple graphic lines begin their developmental path, but later the lines develop into structures with distinct formal characteristics of living systems, such as plants and various animals. Among the developers of these systems is the developmental biologist Richard Dawkins, author of "The Selfish Gene". These systems are based on "genes" that determine different "form features". It turns out that even a "creature" (line), which begins its journey with a certain "genetic" load, develops in the course of many runs (many "generations") formal features that it did not have at the beginning of the route. The "creatures" that develop in this system are remarkably similar to the creatures that did develop on Earth during biological evolution.

Another system, developed by the artist and software developer Karl Sims (1962 Karl Sims), is a kind of evolutionary habitat for the development of "creatures" built from a kind of slabs and cubes, and which, despite the limitations of these basic building blocks, manage to develop and demonstrate mobility and even behavior that largely resembles their own behavior of real animals, such as birds, crabs, tadpoles, caterpillars, lizards and even a kind of dog that "wags their tails in circumstances that seem appropriate to them".

A computer system was built at the institute in Santa Fe, where programs are written with the aim of developing the most prominent feature of living systems - self-replication. The "capricious" interventions of the scientists in the software race, within the framework of the principles of evolutionary programming, sometimes result in the development of additional "creatures" in the population of certain "creatures" that function as "parasites" that take advantage of the first "creatures" (which are actually nothing more than a few lines of software), For their self-replication - which they are unable to do on their own - just like viruses in the "real world". Later, "creatures" develop that are able to repel the parasites, which causes the development of "parasites" that are also able to trick the new mutation. This is how an evolutionary arms race develops, typical of biological systems, and in a relatively short time the computerized space is flooded with dozens of different species of "creatures", which eat each other, exploit each other, grow old, die, give birth to offspring and more, as is customary in the living world.

But the resemblance to real animals that exist on Earth is not a mandatory condition in relation to attempts to create "any" life (for example, within a computerized system). The inability to clearly know the chemical composition of the primordial soup led John McCaskill (1957) and Walter Fontana (1960) to open a new field of research, which they called "artificial chemistry" or "algorithmic chemistry". This is a system that, instead of trying to get to the truth about the processes of the formation of life on Earth, can suggest to scientists to create new worlds: to establish new chemistry laws, new materials and new rules according to which these theoretical atoms and molecules connect to each other or break chemical bonds. It is a very complex system, capable of controlling a large number of virtual atoms and molecules and managing their reactions with each other and with the environment. In this way, new "materials" are created that may lead to the development of properties that, according to our perception, are related to life.

In the theoretical worlds of algorithmic chemistry, there can, of course, also exist evolutionary principles that differ from those we know on Earth. Fontana, who showed a willingness to allow and test different theoretical evolutions, showed courage and took a risk that could have ended his scientific career in the real world. He gave up a permanent position at the university - a clearly anti-evolutionary act - and instead accepted a limited position for six years at the Institute for the Study of Complexity in Santa Fe to concentrate on exploring the possibilities of creating artificial life. In the end he "landed" safely at Harvard.

Change the rules of the game

In the last decade - mainly in Japan and Germany - there has been intensive research in artificial chemistry, only part of which concerns the question of the development of life and its definition. The great complexity of artificial chemistry systems allows them to be used to solve various and very complex problems, such as, for example, controlling a small robot. Tim Hutton (1975), from the research division of Microsoft Research, built in this way a complete, new artificial environment, based on simple chemistry and physics - which support processes that lead to the spontaneous development of molecules, which are able to accelerate the "chemical" reactions they carry out - and later also reproduce themselves.

Houghton sets the laws in his world: what is the size of the arena where the process takes place, what "atoms" and "molecules" will be found in it and in what quantity, and what laws will govern the processes of joining these atoms and molecules to others and the processes of disintegration of the products of previous "chemical reactions". In the experiments he runs, the system begins its journey from a "primitive soup", and the self-replication process that develops in it, without a guiding hand, is reminiscent of the DNA replication process. During the "generations" more than once in his system "offspring" developed which have distinct evolutionary and survival advantages over their "parents" - just like in the real world. This system may also be used for a variety of applications unrelated to the origin of life, including calculations of changes in economic systems, population studies in sociology and even linguistics.

Some argue that the mathematical models underlying these systems represent life for everything. Others believe that the models are not alive, although they perform an ever-improving simulation of life. Both of them agree that the research in this field is still far from saying their last word.
In any case, one of the interesting phenomena observed in the complex computerized system, in which the self-replicating "creatures" develop, shows that under certain conditions, even the seemingly victorious "creatures" are not eternal. No one is infallible. They too are defeated in the end and disappear from the stage of history (the background), when they make way for new "stars". Some consider this feature of theirs (which can be seen as "death") the strongest proof of the certain parallel between them and between "real" living systems.

Life as a work of art

The new spirit of synthesis and multidisciplinarity, which characterizes both the sciences and the arts at the beginning of the 21st century, leads artists and scientists to experiments in developing interactions between humans and machines, and sometimes also between humans and virtual entities that simulate life. Carl Sims, one of the first artists to use the principles of artificial life in their work, invited visitors to choose forms that later developed as a result of "mutations". In this work, which he called "genetic images", a mixture of the choice and preference of humans (the visitors to the exhibition who interacted with the work) was created with a kind of "artificial genetics".

Between 1993 and 1994, Christa Sommerer (1964 Christa Sommerer) and Laurent Mignonneau (1967 Mignonneau Laurent) presented an interactive virtual work, which they called A-Volve. A computerized system that offered visitors to create artificial creatures, come into contact with them and observe their development and the way in which they react to human contact. The technology underlying the creation was none other than Thomas Ray's Tierra system, described in the previous pages.

A-Volve by Shel Sommerer and Mino is actually a kind of interactive computerized environment, which includes a glass pool of water whose hand movements of the users, who are standing around it, affect - by means of digital cameras and sensors - what is happening on a computer screen. The virtual creatures are created spontaneously and randomly, but also as a "creative act" of the visitors. The creatures that are "born into the pool of water" begin a complex course of development while competing and fighting, with the attributes and tools at their disposal being energy levels, speed of movement, reproduction rate and lifespan. The "creatures" are able to pass on their "genes" from generation to generation, during natural selection and evolution. The algorithms developed by Sommerer and Mino ensure that the creatures will exhibit natural and smooth movements similar to animals. It is not about the choice of the critics-creators between "models" of pre-planned creatures. The creatures are created - each time in a new way - through interactions with the visitors and among themselves. This is how a variety of creatures is created that enables evolutionary processes similar to those that occur in nature.

The visitor, who wishes to "create" an artificial creature, does so by pressing the interface "buttons" of the touch screen, and by "drawing" on the background. Sommerer's and Mino's algorithms (based on Ray's system) calculate the XNUMXD structure of the creatures and their movement in the water. The creature's movement and behavior are also affected at this stage by the visitor's preferences, as they are expressed through the touch screen. To a certain extent, the creature's behavior in space is an expression of the form, and the form is an expression of adaptation to the environment. That is, the shape of the creature and the way it moves are related to each other, so that its ability to move defines its suitability for life in the pond.

The fittest creature will survive longer and be able to mate and reproduce. The creatures compete with each other for energy resources, and the competition in this case may lead to situations where one creature preys on another creature to get its energy. The creatures are also affected by the visitors: if a visitor tries to catch a creature (by wading his hand in the pool water), the creature (on the computer screen) will try to escape, and if caught it will freeze in place. Thus the visitor can influence evolution, for example by protecting a creature from its predator.

When two surviving creatures meet and mate, they can create offspring and thus a new creature is born. The offspring will carry the "genes" of its parents. The virtual reproduction mechanism includes mechanisms of mutations and "gene" exchange.

In the first stage of the creation process, the visitors are asked to draw by touching a finger on a touch screen, a two-dimensional side view (profile) as well as a cross-section of the creature they wish to create. These two actually define the three-dimensional shape of the creature.

The newborn is allotted only one minute of "life" (as a result of the limitation of computing power that existed in the early 90s for Silicon Graphics workstations, on which the system "ran"). During this time he must eat, survive, mate and produce offspring. Otherwise, he will die (as a result of running out of time allotted to him, starving for energy or being eaten by another creature) - without being able to pass on his "genes" to the next generation. At any given moment no more than 20 creatures can "live" in the system at the same time.

The Austrian Sommerer and the French Mino met at the Complexity Research Institute in Santa Fe. From there they moved together to the Advanced Communication Research Institute in Japan. They are currently professors of art and head the department of cultural interfaces at the Institute of Communication at the University of Art and Design in Linz, Austria. They chose the path of "art as a process", while consciously renouncing their role as artists in the creative process - or at least questioning the artist's role in the "system". Like James Kidder in Sturgeon's story "God in Little Enfin", they "just" lay down the initial foundations, and then let the system (which includes the visitors to the exhibition) operate independently, create and reinvent itself each time, cope, fail, succeed and develop. In this sense, Sumerer and his men, in their renunciation, did not do an act of humility, but the opposite: they put themselves on the level of the "God in the smallest of Enfin".
The young Israeli artist, Lior Ben-Gai (1984 Lior Ben-Gai), examines in his works from 2013-2016 the definitions and limits of the concept of "life". He does this through the creation and programming of complex systems of artificial life that start from the process of the formation of a single creature, and reach questions of complex behavior of "colonies" similar to those of micro-organisms. The coordination in which these colonies work raises questions about the evolution of multicellular organisms.

the artificial man

If we can create artificial life, because then knowledge gives that other creatures, more advanced than us, can do it too. And if so, perhaps, in fact, we ourselves are those artificial beings created by an advanced culture, which observes us from a distance and studies us, as James Kidder, Sturgeon's hero in "El Bezair Anfin", watched the "Neuters" he created and studied their guests ? Nick Bostrom from the University of Oxford made exactly this claim, which led to a discussion from which it appears that in fact, we can never know if we are "artificial beings" inhabiting some kind of simulation or if we "really" exist.

Thomas Metzinger (Thomas Metzinger 1958), from the University of Mainz in Germany, continued this line of thought and proposed to be satisfied in the first stage with a slightly easier question: where exactly does the feeling of "I exist" that characterizes the vast majority of humans (with the exception of those suffering from Cotter syndrome, which is expressed by the fact that they are convinced that they do not exist). A reversal of this phenomenon may, according to David Chalmers (David Chalmers 1966) of New York University, lead to the view that every object and every material (toys, rocks) has a certain degree of life and awareness.