Zhou 祎 Dialogue Kevin Kelly: Will artificial intelligence kill humans?

In mid-December, I had an in-depth dialogue with the Internet prophet, former editor of Wired magazine, and Kevin Kelly (KK), author of Runaway. In the process of dialogue, KK says AI is the technology that is going to be a big line over the next 20 years, but AI will not replace humans. KK said that humans are destroyed by robots is a very typical Hollywood scene, the first robot needs to know how to eliminate the concept, and the robot to eliminate human beings do not have any benefits, so this idea is not reasonable.

Here is a transcript of our dialogue:

Zhou Hongyi: in your book, you think that technology is actually the first person to create, but slowly the technical system becomes an independent system that can evolve independently of the six major systems of biology. In your imagination, technology will develop to a very high degree. So in the next 5-10 years, will technology be as depicted in the book "Singularity near", so that human civilization can get a huge leap?

KK: I think technology is the most promising to improve AI. AI develops very fast. In the next 5-10 years, it will make great progress. There is no doubt that it can bring us into a new height. I think it can be as great as the Internet. Just like the Internet, which started in the 80 's, we have been through 20 years and have undergone tremendous changes, especially business and culture. AI will also experience such a process. I think Ai has the potential to change the world again within 20 years, such as creating 10,000 new businesses, changing culture, and guiding our evolution. I think AI is the technology that will be in the pipeline for the next 20 years.

Zhou Hongyi: There is a movie, I do not know you have seen no, called "Super Body", in this movie, it tells a story, is the human brain utilization rate seems to be only 5%, in your book mentioned a lot of technical changes but did not mention the human brain itself ... is not to do more imagination. It means that when the human brain's utilization rate reaches 30% or 50% or higher, the evolution of humans will have a greater impact. Do you think it's possible in the future? What do you think of the increase in the use of the human brain as described in this film?

KK: I think the idea that the brain of the "super Body" is being explored in this movie is actually not right for 5%. But even so, you can imagine that humans are making themselves smarter in some way, and that's the power of AI. We need not even smarter humans, but ideas that are different from the way people think. This is the power of the computer, it will not be as human thinking as the constant error. It is perfect, because it is more than just imitation of human beings.

If you have a computer, you will find that it does not always forget like humans. It's strong because it's thinking in different ways. When we create artificial intelligence, we are not only Superman in the creation of artificial intelligence, we want something different. Of course we want it to be brilliant, but to think in a different way. Just like an old Apple advertising word says, "Think different", this is a kind of ability. So what we want is a lot of ideas that are different from human thinking, and I think they look like humans, but they're smarter. We want them to think in different ways, and that's what we want to figure out, just like quarks and gravity. We need brains with different ideas to think about these issues, so I don't think we'll try to make humans smarter in the future. But we would like to think in a different way, and that would be smarter.

Zhou Hongyi: as you mentioned in your book, humans were just one of the species on Earth, but 100,000 years ago, after Neanters, the intellect of modern people made a leap, so humans became the most powerful species on earth, and then humans quickly migrated to all continents, quickly destroying other species. , because humans have evolved much faster than other species. But if we look at recent human 3,000 years or the last 100 years of history, "technical elements", get a rapid development. But the evolution of human self has not seemed to have made any special progress in these 3,000 years. For example, we are actually like primitive tribes. If there is a shortage of energy, each country must be like the original tribe to rob Prey will fight, but with missiles, rockets, machine guns to replace the traditional machete spear. Human beings do not have the same development as technology. So according to this, will the future machine civilization be eliminated from human civilization?

KK: robots will destroy humans. The idea is a very typical Hollywood scene, but I don't believe this. Because this requires the robot to know how to eliminate the concept, it is very intelligent for artificial intelligence to eliminate this thing, and it needs many different, intelligent ideas like robots to work together. and human beings can think in different ways, we and robots compared to the human is valuable, there are different ways of thinking. I don't think this is a reasonable idea to eliminate human beings from robots, so I don't believe it.

Zhou Hongyi: Now there are two types of artificial intelligence, such as artificial AI, such as deep learning algorithms, the use of big data, it shows the behavior, such as the machine is like intelligence, but the machine does not produce consciousness, and another is what you describe in the book, If the entire internet on the earth as a neural network, each computer can be regarded as a neuron, then in the future as the number of connected computers, smart devices are becoming a huge internet, because of the increase in the size of the node, it may generate machine self-awareness, And then make the machine produce real human-like intelligence, what direction do you think the future of AI is going to be?

KK: in the near future, within five years, ten years or 20 years, the first AI will become very mechanical and very boring. It will be sold like internet services by Google or Amazon. Because they are unconscious, they are just a variety of intelligent products. You will try to buy the smart products you need and apply them to your questions, customers, or whatever else you want. So the process is like buying appliances, it's not attractive at all, and there's nothing exciting about it, it's boring.

Another kind of AI can be tricky when conscious and aware of itself, and I think we should try to prevent this from happening. Because when it has consciousness, it is no longer practical. So it's better not to use something that you always want to know about yourself. But I think if we create a global internet with a very smooth connection, artificial intelligence may have this idea, the idea that we don't understand, even if it's unconscious. It's a novel idea, a completely different height, and it's too difficult for us to control it because it's very high. So it would be a completely different challenge for us. But I think the vast majority of 99% of AI will only be industrial AI service-type appliances, but that will also be very powerful. Electrical appliances, transformation are civilized, and artificial intelligence will be so.

Zhou 祎 Dialogue Kevin Kelly: Will artificial intelligence kill humans?