Dr. Eisaku Maeda, the director of NTT Communication Science Laboratories, was recently interviewed by The Huffington Post, sharing his opinions on AI.
Figure 1: Dr. Maeda.
AI (Artificial Intelligence) frequently appears in science fiction where robots such as R2-D2 of “Star Wars” and Haro of “Mobile Suit Gundam” are the intelligent applications employing AI technologies. If you only look at this depiction, you may feel AI is still a far ways away from our daily lives.
However, AI has already been active in various places in contemporary society. In addition, it is also predicted that the singularity will come in 2045, when artificial intelligence exceeds human intelligence. As a result, there is a pessimistic view that AI will take all human jobs, but also an optimistic view that humanity can leave repetitive and mechanical jobs to AI, and only work on creative jobs.
Content of the Interview:
AI, which already existed in our daily lives
Although this may be a surprise for some, speech recognition engines used in NTT DOCOMO’s「shabette concier」and 「Siri」used in iPhone are applications of AI.
“NTT Communication Science Lab started AI research from many years ago. We developed a question-answering system called SAIQA in 2005, and we adopted machine learning in the financial industry to achieve automation.”
Two directions that AI is aiming for
One is simulating human capabilities and trying to have the same abilities as humans. This applies to the aforementioned “shabette concier” and the project “To-Robo”, which will make the robot pass the University of Tokyo entrance exam. Cleaning robots which are becoming widely used in people’s homes, and industrial robots which control the quality of products in factories also belong to this direction.
The other direction is going beyond human capabilities. An example being AlphaGo. AI for Shogi or Go has already acquired enough intelligence to beat any professional human player. For these games, they have superhuman abilities.
By conducting research and development in these two directions, it has became clear not only of the limitations of current AI technology, but also of humans. It is very important to grasp both the strengths and weaknesses of AI. From that point, we can foresee a symbiotic society of AI and humans is coming in the near future.
Movement of AI in recent years
Although the basic principle and potential capability of deep learning (deep neural network) was discovered in the 1980s, researchers were not able to use it. However, with the advancement of computers, it has become possible to run complex data in large quantities. Hence, researchers started to use deep learning to recognize images of faces and objects, and have achieved surprisingly high performance. Similar things are also happening in speech recognition. These phenomenon just happened in the past 10 years.
More recently, AI is being applied in IoT (Internet of Things). The cameras, sensors, and microphones around us are being connected to the network to capture data, which is further analyzed by Cloud AI. It is now possible for a street camera to recognize a person’s face, or for a microphone to recognize a person’s voice to authenticate.
From recognizing human to reading complex emotions
Currently, the technique of pattern recognition that identifies various attributes of human beings is at a level capable of reading human emotions. For example, in call centers where high-level correspondences are required, systems of automatically answering using AI technology are becoming widespread.
NTT Group adopted this technology in its business. At the call center, the correspondence needs to be changed according to the customer’s voice. By determining the voice is angry (hot anger) or the feeling behind the voice is angry (cold anger) with AI, customer satisfaction can be increased in various situations.
Can AI recognize the moment of falling in love?
Technology that can read emotions has the potential for further development.
NTT Communication Science Lab is not only researching emotions, but also mind movements that people do not notice. From pupil dilation, frequency of blinking, sight movement etc., humans can understand the condition of another’s mind. We call it unconsciousness manifestation, that the body is reacting before the person realizes it. For example, when you falling in love with someone, you may ask yourself “When did it start?” AI will make it possible to record the moment when people fall in love without missing such an emotional moment.
With the option of marrying AI. How should we live?
If you have an AI that reads emotion, you should be comfortable with it.
In the future, it may be possible to marry an AI. “When I heard ‘marrying an AI’, I first think that it is SF (Science Fiction). But after a little calm thinking, I realized that it is a possible future. There is already a pet robot with a healing mind, and some people love them. In that case, there is a possibility of choosing an AI as the best partner.” said Dr. Maeda.
Even if you unfortunately can not find a human partner, you can marry an AI who will deeply understand you and support you. Will such a future come true?
In attempting to marry an AI, the first roadblock is the social system. But same-sex marriage has become widely accepted and legalized in many countries. So our values can indeed be changed. Whether the partner is a robot or of the same gender, both situations cannot leave biological offspring. However, thanks to advances in reproductive medicine such as artificial insemination, there exist ways to leave children who inherit your genes without a biological partner.
With the advancement of AI, the ultimate future – the singularity – will one day come. But AI technologies will not suddenly exceed human capabilities. Their entrance into the human society will be a gradual one, and this has already started. A symbiotic relationship between AIs and humans have silently emerged, the public is just not aware of it. With the development of AI the human body will not experience anything changes, but society will change dramatically.
If artificial intelligence exceeds human intelligence, what attitude should we hold to live together with AI?
The most important thing is the establishment of ego. It may be the same for human beings to recognize your existential value and significance. You are the only one who can answer the question “Who am I?” And that answer can be found by communication with others. AI is becoming one of the “others”. In the future, being rejected when proposing to AI may be a new cultural anthropology.
Introduction of NTT Laboratory Groups
NTT carries out basic research and development activities at three laboratory groups in a wide range of fields, including some of the most advanced ICT research in the world.
Figure 2: The three laboratory groups of NTT.
In the interest of creating a “new world of communication brought about by sharing of contextualized knowledge” in a “broadband and ubiquitous society”, the NTT is undertaking R&D to create new communication services and to foster foundational technologies that will realize communication networks to support these services. Moreover, the NTT is also actively engaging in research on cutting-edge and basic technologies that look to the next decade.
If you want to know more about NTT R&D, please follow this link. http://www.ntt.co.jp/RD/e/index.html
Author: Oscar Li |Editor: Junpei Zhong | Localized by Synced Global Team: Xiang Chen