By Tang Guhan
Self-aware robots are not going to take over the world, and they will create productivity in the future, said Hod Lipson, an American Robotics Engineer and the director of Columbia University's Creative Machines Lab.
"We have seen incredible achievements in the field of Artificial Intelligence (AI). If we traced back to six years ago, computers could not tell the difference between a dog and a cat reliably," said Lipson in an exclusive interview with the China Fortune Media Group under the Xinhua News Agency, "but now they can do better than average people regarding their capabilities to understand images, video and audio clips. These have all been happening within two or three years."
However, the area where robots are not good at is language, Lipson noted. "they can understand the words of what you say but not the implications behind the words. For example, computers can understand what kind of toppings you want when you order a pizza but cannot understand when you have a philosophical conversation."
"But this is the area where we could see breakthroughs in the next decade," he said.
"Everyone around me is talking about the AI around computer clouding, virtual and cyber but what I believe is there will be breakthroughs in the physical area to combine AI and analytic capabilities in the next five to ten years," said Shih-Fu Chang, the Senior Vice Dean of the Fu Foundation School of Engineering and Applied Science from Columbia University.
"Robots are all a hard metallic surface; it is not soft material muscle. We can design a physical device that has the manipulation capability using the AI model behind it." He explained.
AI can speak, can read emotions, create art and music, but can they "feel" like a human?
Lipson is an expert focusing on evolutionary robotics. He has built self-aware robots who can learn by themselves and have self-consciousness.
The robot can sense its body, but it is blind. In the beginning, it does not know its shape, whether it is a snake or a spider, but after four days of learning, it learned that it has four legs and gradually created a self-image to know how to walk.
"Not because it was programmed by a computer, nor did it do machine learning, it learned through its imagination, just like humans can imagine themselves."
Why will it matter to the real economy?
Professor Lipson noted a robot could recover from damage. It is resilient and can learn a lot faster without data. "Imagine you have a manufacturing plant that can stimulate itself. If something goes wrong, it will spot it by itself. Meanwhile, it could learn new things without being taught."
However, this is a very nascent field where academia is working on, he added.
Here comes another question, will robots stimulated by AI technology be free from human control in the future?
Lipson emphasized that it is up to us humans to decide whether we want to take that route. "To me, the question is not what AI will do to people but what people will do to people using AI, as an educational institution, we will not only study how to develop AI but also having discussions on where technology is going in the future."
