"My appearance is a lovely girl, but please don't be confused by my appearance, because I have a strong learning ability and can do many things, such as chatting with you, helping you do housework..." Sweet voice comes from a heavy makeup "beauty" with short hair and silver armor, named Xiao Qi. On April 25, when the 2024 Zhongguancun Forum opened in Beijing, this bionic robot stationed at the front desk to serve guests became one of the busiest "people" in the whole audience.
Of course, not only Xiaoqi. Coffee robots with intelligent "iron arms" are used to make coffee for guests; It is the digital person anchor who broadcasts the news for the audience... The most direct impression of the Red Star journalists who went to Beijing to attend the conference is that they "entered the world of human-computer cooperation in science fiction movies in advance".
In fact, the penetration of AI into social life has appeared in all aspects. In particular, the advent of generative AI such as ChatGPT and Sora has made people believe that all walks of life will undergo disruptive changes. So this year's Zhongguancun Forum has AI everywhere, and everyone talks about AI.
Ten years is short, but it is long enough to put it on the axis of future development of AI. How smart will robots be in 10 years? How will it enter into life? Red Star journalists talked with several head experts on the forum to discuss how AI will change the world.
▲ Bionic robot at Zhongguancun Forum future
Like owning a smart phone, to have a talented assistant
In fact, the technology of generative AI has emerged in the past five years, but in the past two years alone, it has brought more wisdom than any genius. "Two years ago, the level of the AI big model could probably reach the level of ordinary people, that is, 34 out of 100 questions could be answered correctly." Li Kaifu, chairman and CEO of Innovation Works, said, but after two years of progress, the best company's big model can now answer 90% of the questions. "It is conceivable that given another two years, the AI big model will have a great opportunity."
▲ Li Kaifu's Speech In the PC era, a computer is placed on every desk; In the era of mobile Internet, people can carry a mobile phone that can connect information and people. What about the AI era? Li Kaifu said: "In the AI era, a talented assistant of IQ300 will be placed beside you to help you think about problems and solve problems." When such a genius appears, productivity will be improved unprecedentedly.
At present, people tend to imagine the application of large models that may appear around them as a chat robot. Li Kaifu predicted that in the next 10 years or so, the chat robot will evolve into a super assistant with full mode understanding ability, able to understand the world in depth, be able to process video, pictures, audio and other information, and also be able to move flexibly and take things. "It will enter all walks of life and become a real robot coming out of science fiction."
▲ The digital human host of Zhongguancun Forum So back to the present, when AI big models have not been widely applied, should the domestic AI industry focus on general big models or vertical big models? public opinions are divergent. When a Red Star reporter asked Kai Fu Lee, he firmly replied, "All things must be done and are important."
Only when a hundred flowers blossom, can we better enable thousands of lines of work in the future.
In Kai Fu Lee's vision, the next stage, "We can expect an AI machine." This machine is no longer just generating content, but can start to help people do things. "For example, if you tell him that my wife's birthday is coming, he will not only answer whether you should buy lilies or roses, but can directly help you buy a gift."
AI will be everywhere
But we should develop and manage well
At the opening ceremony of the 2024 Zhongguancun Forum, the embodied AGI (General Artificial Intelligence) system prototype launched by the Beijing General Artificial Intelligence Research Institute was "Tongtong" and selected as the major scientific and technological achievements of the Forum. The image of "Tongtong" is that of a 3-4 year old child. It has a unique "mind", can understand words and understand intentions, and its performance is similar to that of a human 3-4 year old child. In the future, "Tongtong" will enter into life, such as serving people tea and water, providing warm companionship, etc.
Ten years later, the iterated "Tongtong" or "Xiaoqi" people will have more powerful wisdom, and will also penetrate people's lives more deeply. The development trend of artificial intelligence has been overwhelming, so when discussing the future, there will be a topic of "how human and computer coexist".
At the time of interview, the Executive Director of China New Generation Artificial Intelligence Development Strategy Research Institute, the Director of Advanced Computing and Key Software (ICT) Haihe Laboratory Gong Ke, the former president of Nankai University, expressed his opinion to the Red Star News reporter: "AI is not an independent subject, it is an extension tool to expand human brain power and labor capacity, just as we do not need to discuss the coexistence of people and cars."
▲ Gong Ke Gong Ke believes that no matter how smart AI is, it is also a tool, and there is no "AI will invade humans in the future". If software wants to become smart, it still needs to rely on hardware, which ultimately depends on scientific and technological breakthroughs, rather than its own metabolism.
Of course, with the development of technology, generative AI is becoming more and more intelligent, but it is not infallible. Gong Ke said that when artificial intelligence speculates on the answer, it uses probability, which is called joint distribution probability. Give it an input question. It judges which words have the highest probability of combining with the input words, that is, the output it chooses. Since it is judged by probability, it cannot be 100% correct. There are always mistakes. Therefore, people should have the ability to examine the facts, logic, ethics and scientific principles of AI, and understand what factors will affect the distribution of probability. People should be able to find and correct problems. In other words, "if one wants to control the tool he has developed, he must know the internal weakness of the tool."
In Gong Ke's opinion, artificial intelligence can be graded and classified for risk classification. For example, brain computer interface implantation into the brain is a high-risk application. For another example, you can distinguish the age of the audience of some applications to determine who uses them and where they are used. Therefore, the development of AI is unstoppable, but it should be well developed and managed.
Red Star Journalist Wang Yao Photographed and reported
Edit Order
