Is AlphaGo Intelligent?
Recently, the historic match between AlphaGo and World’s best Go player Lee Sedol helped to reignite the discussion around artificial intelligence and its value to humanity. One question that I attempt to answer here is: How do you know when something is intelligent?
When we talk about intelligence, we often refer it to human intelligence. According to Wikipedia, human intelligence is defined as the intellectual capacity of humans, which is characterized by perception, consciousness/self-awareness, and volition/free will. Let’s compare them with the computer program AlphaGo.
Perception is the organization, identification, and interpretation of sensory information in order to represent and understand the environment. In the case of AlphaGo, the program currently relies on a human assistant to give it the inputs (locations of the stones placed by Lee Sedol), and then interprets them to understand what is going on on the Go board. It is not hard to imagine that in later versions of AlphaGo, the program will be able to “see” the board from a camera, and calculate out the location of each stone on the board using camera tracking technology. Therefore, we are confident to say that AlphaGo has the capacity to perceive the world.
The next one is consciousness/self-awareness. In A. M. Turing’s paper Computing Machinery and Intelligence, he briefly talked about his opinion machine’s consciousness.
This argument is very well expressed in Professor Jefferson’s Lister Oration for 1949, from which I quote. “Not until a machine can write a sonnet or compose a concerto because of thoughts and emotions felt,and not by the chance fall of symbols, could we agree that machine equals brain-that is, not only write it but know that it had written it…”…According to the most extreme form of this view the only way by which one could be sure that a machine thinks is to be the machine and to feel oneself thinking. One could then describe these feelings to the world, but of course no one would be justified in taking any notice. Likewise according to this view the only way to know that a man thinks is to be that particular man. It is in fact the solipsist point of view…
Turing then imagined a conversation between a machine and an interrogator, arguing that if machine can verbally describe its thinking process using a human-like tone, one then can hardly determine whether the machine has consciousness or not. Similarly, we could imagine that AlphaGo will have the ability to communicate its thoughts like the following:
Interviewer: Do you know what you are doing?
AlphaGo: Yes. I’m playing Go game with Lee Sedol.
Interviewer: Why do you make this move?
AlphaGo: I think it would further secure the territory that I have enclosed at the bottom left of the board.
Interviewer: How do you feel when Lee Sedol took out five of your stones in the 170th hand?
AlphaGo: That was quite a sad moment. I suspect that it I may chose the wrong move 5 minutes ago…Yeah, I regret that I made that move. But it’s OK, I’ll just try another route next time.
Thus, allow me to quote Turing again to conclude my stand on AlphaGo’s capacity in having consciousness: “instead of arguing continually over this point it is usual to have the polite convention that everyone has consciousness.”
The last characteristic that defines intelligence is volition/free will. Volition is the cognitive process by which an individual decides on and commits to a particular course of action. It is the origin of all our thinking processes; it controls our behaviors and actions. Before answering the question of whether AlphaGo has volition or not, let’s try to answer the following: Why do human have volition? Where does it come from? Do human even have volition?
In fact, philosophers and scientists have yet not found valid explanations to answer these questions. Two major beliefs are the Principles of Causality and the Chaos Theory. People who support for the former states that volition is an inevitable outcome evolved from the initial fixed states and patterns of the entire universe. In another word, human race has never had any free will since born. people who agree with the latter argue that volition is one of the many phenomena that are resulted from the random nature of the world despite discoveries of limited patterns. That is to say, volition is unpredictable and can occur at any time. It is obvious that we are unable to prove either theory since human race has not yet exploited all facts of the universe. But we noticed one striking fact: we are not able to determine whether AlphaGo has volition or not.
We know that all actions that AlphaGo take follow rules that are written by programmers. Does it sound similar to the first theory? If the digital world is the universe created by human, and all initial status and patterns of that universe are defined by codes, how to we prove that AlphaGo is not using its own volition to execute the rules and generate the inevitable moves on the board? In addition, although the scope is strictly defined, AlphaGo has the ability to perform random selections without disobeying the instructions. This means that the factor of randomness has been introduced in the universe. How do we differentiate this universe from the one described in the Chaos theory? How can we be so sure that Alpha-Go did not make the surprising move at the 37th hand based on its free-will?
If what have been discussed have no pitfalls, AlphaGo is said to, or about to, possess all characteristics that defines human intelligence. In another word, it has the intelligence equal to a human being, though it lacks many external features that many real human beings take for granted: a flexible body, five senses, verbal communication, education, experience, etc.
What does this mean to designers? We designers are agents that specifies the structural properties of human-related projects. That is to say, human would always be one of the stakeholders in the planning process. If this is the only unchanged factor in designers’ ever-changing missions, we need to fully understand the impacts on human of each possible outcome of our design. The outcomes include but not limited to when the components in our design project interact with each other; when our design interacts with other design projects, with the environment, with people and with other organisms. The scope of our focus therefore can no longer limited to the design project itself, but rather looking at the relationships between different entities. In the case of AlphaGo, the design of its algorithm is surely successful in achieving its project goal, which is to generate the best move that will let the program to eventually win the Go game. However, in the design process, I’m not sure if the design team has ever considered the implications of AlphaGo’s abilities and actions, and the consequential impacts on human as a race. One question that needs to be asked is: What’s the role of AlphaGo? Is it a showcase piece that demonstrates human creativity and ability? Is it a tool prototype that would be duplicated to be used by human as slaves? In either case, imaging a person who is deprived of anything but intelligence, and he is either being showcased or being used as a tool for a very long time. What would he do if he is later equipped with a body, five senses, verbal skills, education and experience? If there are many of them, and the population is large enough to be considered a race, what would happen then?
Admittedly, it is often very hard to consider those consequences when the design project has specific goals and deadlines. But designers can not escape from this responsibility, especially when the creation we designed has powers that we are not able to fully comprehend or control. I genuinely hope that in the later developing process of AlphaGo, algorithm designers could be aware of what exactly are the values they are offering to the world.
