The question `What is Artificial Intelligence(AI)' is a difficult question for AI researchers to answer.
Japanese Society for Artificial Intelligence has published a series of articles to discuss
the question `What is AI?' in JSAI journal in 2013.
Following is English summary of my article in this series.
Short answers to the question `What is AI?'.
In the series articles to discuss `What is AI', four professors already
tried to give answers
before my article.
The answers were:
Professor Nakashima: Artificially made entity that has intelligence. Or, the field to study the whole of intelligence by trying to make such entities.
Professor Nishida: `Mechanism that has intelligence' of `Mechanism that has mind'.
Professor Mizoguchi: Artificially made something (or system) that behaves intelligently.
Professor Nagao: Systems to ultimately simulate human brain activity.
My (Koichi Hori's) answer is:
Artificially-built new world of intelligence.
Four professors before me used the terms entities, systems, or something.
In contrast, I use the term `world'.
Let's consider the case of `artificial satellite'. When engineers and scientists build
an artificial satellite, they are not aiming at mimicking the moon, which is the
satellite of the earth. Of course, artificial satellites and the moon have many
characteristics in common, but the goal is not to mimic the moon. To achieve
the objectives of observation, communication, scientific experiments, and so on,
not only the satellite itself but also related ground stations and operation systems
are developed and deployed.
In the case of artificial intelligence as well, we are not just aiming at
mimicking the human intelligence, but we are building new worlds of intelligence
where many goals are achieved.
What is the new world of intelligence?
Then, what is the world of intelligence? To consider the problem, let's think of
literature and business. What is literature or what is business?
The novels or poems themselves are not equal to the literature.
The authors, the works, the publishers, the media, the critics, the readers and so on compose
the world of literature. Likewise, the business persons, the merchandise, the market, the rules, the consumers and so on compose the world of business.
The activity of human brain is not equal to intelligence. We should build the whole
world of intelligence.
My answer to the question `What is Artificial Intelligence' is:
artificial intelligence = new worlds of intelligence, which are synthesized artificially,
a world of intelligence = the whole of the elements and the whole of the mutual relationships among the elements,
the elements = human brains, human bodies, tools, problems, solutions, data, information, knowledge, wisdom, values, emotions, languages, machine programs, machine bodies, machine networks, human networks.
Maybe, I can say that Philosophy studies the whole world speculatively and Artificial
Intelligence research builds the whole new world synthetically.
Some readers may think my definition is too broad. My definition of artificial intelligence
could include jurisprudence or economics. But, as a matter of fact, the mixture
of juristic studies and AI studies, the mixture of economics studies and AI studies,
or the mixture of ethics studies and AI studies
are beginning to happen.
AI researchers are now requested to consider the whole world. We cannot be just
Artificial Intelligence vs. Intelligence Amplifier
I myself is more interested in making humans more intelligent than making machines
more intelligent. The studies on the systems to make humans more intelligent have
been called intelligence amplifiers.
At first, since people thought that the intelligence amplifiers were just tools,
people did not think deeply the intelligence amplifiers mean for us.
However, we now know that the intelligence amplifiers can change the whole world of
intelligence. The persons with the tools are not same as the persons without the
Moreover, the tools to enhance the people are now beginning to be connected to
each other through networks.
We should carefully study what should be the appropriate goals achieved by
enhancing the humans with the new tools, that is, intelligence amplifiers.
My definition of artificial intelligence will naturally lead to the problem of
AI ethics, because AI research changes the whole world.
When we discuss the AI ethics, it it not sufficient to discuss just the ethical problems
of machine intelligence. As I mentioned above, the whole world of intelligence
are composed of many kinds of elements and the relationships among the elements.
Ethical problems will emerge from the whole of the relationships.
This is a problem of `complex system'.
AI is now requiring new paradigms of ethics studies.
Fortunately, we, AI researchers, are now working with the ethicists to study this