Google held the annual Google I/O conference, which showed the magical power of Google's newly developed artificial intelligience solftware Google Assistant. For example, a user wants to cut her hair, then Google assistant directly helps the user to make an appointment for the barber shop. The Google assistant did not show any lag or logical error in the real-life conversation with barbershop. Miracly, the barbershop did not notice that he was having a conversation with AI at the moment. Besides, the AI can also help people to book restaurants, theaters, travel agencies, and learn to chat with people like a small / girlfriend / old friend, while achieving multi-line processing.
The features demonstrated by Google's assistants mean a further upgrade of artificial intelligence and a milestone. If the staff of the barbershop are not aware of the conversation with the AI, it proves that the Google assistant passed the Turing test and has the same or similar intelligence as the real person. Even so, it is not the AI'S anti-day or subversion of human beings because this is only a further breathrough in artificial speech and logic, and it cannot fully possess the same wisdom or intelligence as human beings like design, manufacturing, and creation etc.
Uber's self- driving car accident in the United states illustrates the low energy of AI. At 10 Pm on March 18, a driverless Uber car crashed into Arizona's 47-year-old pedestrian Elaine Herzberg. The preliminary investigation now shows that the Uber self-driving car has detected pedestrian before the accident, but Uber autopilot software(AI) did not respond. Part of the reason is that the technology is adjusted to ignore objects that are indentified as "false positives", such as plastic bags on the road. That is to say, although the self-driving AI has found pedestrains, it has mistakenly regarded pedestrans as plastic bags, which is the significant difference with human mind—AI'S deep learning ability and human comprehensive intelligence.
Human perception and cognition are based on the comprehensive characters of external affairs and comprehensive analysis to quickly draw conclusions and take measures. Sometimes it is called intuition. However, this perception is based on comprehensive perception and analysis. For living objects (humans and animals), human cognition depends on not only apperance, body shape, but also dynamic (motion), three-dimension, odor, direction etc, even the direction can be distinguished as up and down, opposition, same and horizontal direction. In the event of an accident, Herzberg is pushing the bicycle. This sporting feature combined with the characteristics of the person, allowes the driver to quickly determine that there is a person in front and must brake or turn a corner. Unfortunately, even with these identifying features or elements( most importantly the ones walking), the AI driving software doesnot recognize people, but as being inanimate.
The upgraded Google assistant and Uber auto-driving car accident once again raised a foundamental problem with artificial intelligence: artificial intelligence is purely biomimetic(human-like) wisdom, independent or distinguished wisdom from people. Some people think that artificial intelligencce is not bionics, but others think that artificial intelligence is the function of imitating human brain, and even if it is not completely bionic, it is the comprehensive intelligence that imitates human brain. strictly speaking, artificial intelligence is emulating the human brain from the terminology of artificial intelligence and comprehensive wisdom. Another development of artificial intelligence is much more illustrative. For example, the Deep Mind team that created AI Alpha Go designed a large artificial neural network structure( artificial intelligence software) that was placed in a virtual reality gaming environment. After intensive learning, the artificial intelligence's ability to navigate the destination in the game maze is beyond that of average person, to reach the professional level—it can find new route or the short route like a mammal.
This principle is developed by mimicking different neural cells in different directions of human brain cells. There are three kinds of cells that brain uses to recognize the direction, namely, positional cells,directional cells, and grid cells. Artificial intelligence has learned the ability of people to recognize directions, and if applied to automatic driving, of course, it can save time, detours and gasoline, however, the ability to avoid accidents does not depend on the cells function with all directions, but brain's other functions—the ability to combine object motion and spatial distance, and to identify the appearance. Artificial intelligencce has long surpassed human capabilities in many aspects of individual intelligence For example, the financial robot developed by Deloitte last year can replace15 finance people, and the robot can finish 360,000 hours of manpower in a matter seconds. In the future, 70% of work including finance, calculation, statistics, low-end technology, and physical work may be replaced or subverted by artificial intelligence, but artificial intelligence is unlikely to replace the comprehensive intelligence of human brain, and to fully cover field concerning the life and health. Of course, logically, there is also the possibility that artificial intelligence completely replace human brain, but all of it is based on artifical intelligence's comprehensive imitation and transcension towards human brain.