The world’s top robotics experts talk about ＂artificial intelligence threat theory＂
"particle robot" system
The robot we see now may be human, or it may be a mechanical arm and a big machine … … Do you dare to imagine that the robot in the future may be composed of countless "cell" robots, not only human beings have autonomous consciousness, but also can change their shapes at will like the robot in Terminator, and even more like human beings, even if "cells" age and die every day, it will not affect the normal body function at all … …
On November 3rd, many authoritative scientists from all over the world gathered at Tencent Science WE Conference in Beijing to explore the future "small universe". At the meeting, the reporter of Yangzi Evening News interviewed Hod Lipson, a researcher of the new particle robot, director of the Creative Machine Laboratory of Columbia University and professor of engineering.
Yangzi Evening News/Zi Niu News reporter Xu Xiaofeng
Explore the future
In the world of grandchildren, machines will have a sense of autonomy
Let the robot take care of itself and manage itself.
Robots have autonomous consciousness and even bite human beings, which has always been a problem that many scientists worry and argue about. The new "particle robot" reported on the cover of Nature magazine on March 20 this year can move collectively like living cells. Simply put, just like the human body is made of cells, this "particle robot" may also be combined into a robot that can be truly autonomous or even "regenerated".
In the future, robots will have autonomous consciousness, but I am not worried. As one of the leaders of the particle robot research team, Professor Hod Lipson thinks that "it is inevitable that there will be such a day. I don’t know whether it will happen in 10 years or 100 years, but I am more sure that our grandchildren will live in a world where machines will have self-awareness."
For the threat theory of autonomous robot, Hod Lipson uses the example of "fire" to express his views — — Fire is very dangerous and powerful. But do humans wish they had never discovered fire? The answer is no. Because with a fire, you can do many things that you couldn’t do before, such as eating cooked food, heating and so on. Robots with autonomous consciousness are powerful and dangerous, and this technology is worth owning. It is important for everyone to realize what this robot can and can’t do, and make sure to use it for good things.
"Just like the face recognition technology that has been criticized by many people for invading privacy, it is a great help to find missing children; The UAV, which may be developed into an attack weapon, is a useful tool to help farmers take care of crops when it is used in agriculture. Therefore, the mirror of AI is just ‘ Enlarged ’ Whether AI develops in a good direction or a bad direction depends on the intentions of developers and users. We have developed and used many machines now, and the number of these machines will become more and more complicated. One day, humans will not be able to directly take care of so many and complicated robots. We must find ways to let robots take care of themselves. "
"particle robot" should "transform" into nanometer
Or you can simulate a "life-like body" with thousands of widgets.
According to Hod Lipson, the so-called "particle robot" system consists of many individual units, with a simple disk shape, a built-in battery, a communication module, a small motor and a uniquely designed mechanical structure. Every part of the particle robot is very simple and cannot move independently. But if the three particles are put together, when the particles are instructed to expand and contract, they will interact with the other two "neighbors", and you will push me and drag me to complete the straight line walk. After adding more particles, more complicated things can be accomplished. A robot composed of a group of particles can move around, transport objects and avoid obstacles under the guidance of light.
The scientific evaluation of particle robot technology developed by Hod will lead to the fourth wave of artificial intelligence, and the machine will probably have the unique creativity of human beings in the past. The great thing about this "particle robot" is that it is just like the cells that make up the human body. Even if 20% of the particles fail, the whole system can still run, which is also a breakthrough in the viability of the system. Hodglipson said that cells are constantly being produced, aging and dying in people every day, but I am still me, and you are still you. Now build a robot, like a human body, which can be composed of these parts (particle robots). Even if it is partially damaged, it can still function normally as a whole like a human body.
This is completely different from our idea of building machines today. An airplane is made up of many parts. If any part fails, the airplane can’t go to the sky. But the robot composed of particle robots is more like an organism, with brains to model and develop itself, and maybe it can be repaired and regenerated in the future.
Hod Lipson told reporters that in the future, "particle robots" will be smaller and will form robots to operate like tens of millions of cells. So using nanotechnology or at least micron technology is the next goal. It is impossible to have a practical scene until every part of the particle robot is small enough.
What are the benefits of "life-like" robots? Hodglipson said that organisms can be 100% recycled. For example, if you eat a food, your body will recycle the useful substances it provides to you and make full use of it. In the future, there may be a physical machine that can realize this cycle, which is composed of millions and tens of millions of small parts and becomes a recyclable ecosystem. The primary application of this robot may appear in the exploration of the universe. For example, when exploring the moon, these robots can change their shape to adapt to and explore the environment.
international cooperation
Among the 10 doctoral students in the team, 4 are from China.
Do China and the United States have different research focuses in the field of artificial intelligence? NO！
Hod Lipson told reporters that there are 10 doctoral students in his research team, including 4 from China, and even more from China, accounting for more than 10 of the master students. There are also many China people among the faculty and staff of our university.
In the field of artificial intelligence and robotics, the number of papers published in China, the United States and Europe is roughly the same. Hod Lipson believes that in the world of artificial intelligence and robotics, the world is really flat, and everyone has a strong sense of sharing, which has greatly promoted international cooperation. Hod Lipson’s laboratory is visited by reporters to take photos and ask questions day and night, and sometimes they ask very difficult questions.
Some people say that the difference between China and the United States in the fields of artificial intelligence and robotics is that China may be more applied in research, while the United States has done more solid basic research. Hod Lipson said that he didn’t think so. Both Chinese and American scientists are involved in research. In his own laboratory, some researchers pay more attention to application and some pay more attention to basic theory. For example, he has a project of his own that is a robot that can "catch" this action. Researchers will have a lot of contacts with the factory to study whether it can catch 100,000 times without failure, and there are people who pay special attention to the creativity and self-awareness of robots. This is a more philosophical thinking. Therefore, in the wider field of artificial intelligence and machinery, practical and theoretical basic research is widely carried out.