news

highlight:

according to media reports on september 29, kai-fu lee, chairman of sinovation ventures and ceo of 01wang, recently participated in the podcast "tools and weapons" hosted by microsoft president brad smith (tools and weapons). in the program, kai-fu lee shared his in-depth insights into artificial intelligence and discussed its role in society, development trends, challenges and future potential. he views artificial intelligence with an optimistic yet realistic attitude, believing that artificial intelligence is both an opportunity and a challenge, and emphasizes that in the era of artificial intelligence, we should pay attention to the meaning of work, the direction of education, and how to use artificial intelligence to solve global problems. . in addition, he will recount a painful personal experience that not only changed the trajectory of his life, but also profoundly affected his understanding of the role of artificial intelligence on the world stage.

the following is a condensed version of the interview:

smith: i want to start with an interesting and unusual project that you recently worked on. unlike many in the tech world, you didn't just write a book about your personal experiences or technology. you also entered the realm of science fiction with a new book, "ai 2041," which uses imagination to 's story paints a picture of the world we may all be living in twenty years from now. what drives you to choose such a way to think and discuss the future of artificial intelligence?

kai-fu lee: i have always been deeply inspired by science fiction, and it has accompanied me throughout my life. when i was working on speech recognition, the holodeck from star trek was something i envisioned. many technical experts in the field of artificial intelligence, such as rodney brooks of mit, were also inspired by science fiction and entered this field. for those of us working on artificial intelligence, the influence of science fiction cannot be underestimated, as technologists and engineers sometimes lack the imagination to weave technology into stories.

artificial intelligence is an extremely important technical field, and i am eager to make it understandable to more people. however, artificial intelligence is often viewed as a complex and esoteric science that seems out of reach for many. so i think science fiction can play a dual role: on the one hand, it can inspire those working in artificial intelligence to envision the brilliant futures they can help create; explain artificial intelligence to the public in a way that is not intimidating, engaging, and maybe even entertaining.

it is based on this idea that i invited the outstanding science fiction writer chen qiufan to collaborate with me and co-write this book. the technologies covered in the book are all technologies that i believe will be possible within the next twenty years. through his brushstrokes, these techniques are transformed into compelling stories. after each story, i wrote a commentary explaining in detail how each technology works, the positive impacts it can bring, the problems it may raise, and how we can address those challenges.

01 the development of artificial intelligence: from dream to reality

smith: in your first book, "ai superpowers," you looked back on some of your past. of particular interest in the book is an application form you filled out in 1983 when applying for a phd program in computer science. in that table, your description of artificial intelligence is impressive. you described it as "exploring the source of intelligent possibilities" and said, "this is the ultimate journey of human self-understanding. i am eager to get involved in this emerging and promising scientific field." you have witnessed the vigorous development of the field of artificial intelligence. . when you look back at the development of artificial intelligence and compare it with your expectations in 1983, do you think the reality meets your original expectations?

kai-fu lee: the development process of artificial intelligence is indeed full of ups and downs. it has experienced many stagnations and winters, but now we have finally realized my original vision for artificial intelligence. artificial intelligence has become a pervasive technology that is widely used in various industries and creates huge value. machine learning algorithms have exceeded our expectations, not only defeating human chess players but also demonstrating extraordinary abilities in areas such as medical diagnosis, tasks we once thought required human intelligence. these achievements allow us to focus on what artificial intelligence cannot do, which may lead us to a deeper understanding of the mystery of human thinking, or lead us to a new era of superintelligence. so, i think we're ready to take that final step. although this process took 40 years, we have basically reached our destination.

smith: if you were walking down the street and you met someone who was asleep 40 years ago, and he heard that you were working on something called artificial intelligence, how would you describe to him what artificial intelligence today actually is?

kai-fu lee: i would describe artificial intelligence as a completely different way of thinking. we use mathematics and massive amounts of data to learn, make decisions, predict, and build systems that exhibit extraordinary power and intelligence. this is different from the human brain because if you have a task that is relatively single domain and has a lot of data to support the final decision or prediction, then ai can far outperform humans every time. but if it’s something that requires analytical skills, common sense, creativity, or emotion, self-awareness, love, and empathy, then ai will be incapable of doing so.

02 the impact of ai on humans: occupations are redistributed

smith: one of the key moments you mentioned was the 2017 alphago game (note: ke jie’s three chess battles with alphago ended in ke jie’s complete defeat). what exactly happened about that game? does it reveal that artificial intelligence will have a significant impact on human work?

kai-fu lee: alphago’s victory marked the advent of the era of artificial intelligence. this event awakened the world’s awareness of the potential of artificial intelligence. go was once considered an extremely complex game that required a lifetime of effort to master. it required not only tactical thinking, but also strategic layout, and even included the philosophy of zen buddhism. in the field of artificial intelligence, many people had believed that the complexity of go was beyond the capabilities of existing algorithms. however, alphago not only defeated the top players in south korea, but also defeated the world champion in china. this achievement completely subverted people's perception.

this incident raises profound questions on many levels. for example, dr. kissinger mentioned in his book that alphago’s victory was a key node in his realization of how artificial intelligence will shape future diplomacy. in china, people realize that this game invented by the chinese has now been conquered by an english-speaking artificial intelligence product, which has inspired the chinese government and enterprises to pay attention to the development of artificial intelligence. alphago's achievement is equivalent to the "sputnik moment" in the technology world, which originated from the successful launch of the soviet union's first artificial satellite in human history, "sputnik 1" in 1957. at the time, this event shocked it exposed the united states' lagging behind in space technology, inspired the united states to invest heavily in the space program in the following decades, and eventually took the lead in the space race), reminding the world that any hope of remaining competitive. all countries and enterprises must pay attention to the development of artificial intelligence.

my book artificial intelligence superpowers mentions this, comparing it to the awakening of the american space race. the field of artificial intelligence has made tremendous progress in the past four or five years. since the essence of artificial intelligence is to imitate human intelligence, its impact on human work is inevitable. once artificial intelligence can perform certain tasks, we will naturally think about whether we can use these software algorithms at almost no additional cost to replace human labor, thereby saving costs.

smith: in the technology world, there are many people trying to describe the future, and you are one of the very few, maybe even the only one, who helps people imagine the future through science fiction. among the many fascinating stories in ai 2041, do you have a particular favorite?

kai-fu lee: i prefer the first story in "ai 2041", which depicts what is happening now. the story is set in india and tells the story of an insurance company that accidentally reinstates caste-like class divisions by controlling social and business applications, and even interferes with the heroine's love life. what’s interesting about this story is that it makes the seemingly boring topic of artificial intelligence in insurance fascinating, while pointing out that ai applications can have unforeseen negative impacts, even if they are aligned with the interests of users. furthermore, it emphasizes that the challenges and opportunities presented by ai are global, and as other stories in the book illustrate, different regions face common problems with ai technology.

smith: in the chapter “job savers,” which discusses the impact of artificial intelligence on employment, you open up the discussion with a story about creating job-redistribution businesses. in the book, what is meant by "career redistributor"?

kai-fu lee: “occupational redistributors” refer to workers in projects set up by the government in the context of artificial intelligence gradually replacing specific occupations. they are responsible for helping workers displaced by automation technology find new jobs and provide retraining services when necessary. this concept emphasizes the profound impact of artificial intelligence on the job market and society’s need for adaptability in the transformation of the workforce.

smith: some of the content in the book is very enlightening and makes people think about the changes that their own work may face in the future. the three abilities you mentioned—creativity, empathy, and dexterity—are things that ai may struggle to master before 2041. why do you think ai might fall short in replacing jobs that require these capabilities?

kai-fu lee: when i talk about creativity, i mean the ability to think across different domains and the ability to think strategically, analytically, and come up with innovative and non-traditional ideas or solutions. these capabilities are currently lacking in artificial intelligence, because artificial intelligence requires a clear objective function, a goal to pursue, and content to learn, all of which are set by human programmers. therefore, humans set goals. in these situations, humans are not only creative in goal setting but also able to think outside the box and come up with never-before-seen concepts. so i think this is an area that artificial intelligence currently cannot achieve and may still be difficult to solve in the next 20 years.

the second is empathy, which is the indescribable feeling of connection, love, or empathy between people. as a human phenomenon, it remains difficult to explain. where is it in our brain? how do we simulate it, can the computer sense it? so again we know nothing because ai is still quantitative and computational, it has no feelings. if you turn off an ai program, it stops running. it doesn't feel sad or bad. it wouldn't be happy if it beat the world go champion. so it has no feeling and we don't know how to program it. we don't even know where to start.

the third is dexterity, an area where we are likely to continue to make incremental progress. this has to do with our years of evolution and our innate hand-eye coordination. so, we combine fine motor skills with thinking and problem solving, which we call dexterity. ai is getting better at this, but there are still many tasks that are far behind. so i think artificial intelligence will gradually conquer this field, but it will definitely not be completed within 20 years.

03 the impact of artificial intelligence on the economies of developing countries

smith: in "artificial intelligence superpowers," you specifically mentioned the possible impact of artificial intelligence on the economy and employment of developing countries. how to evaluate the current potential significance of artificial intelligence to the economies of developing countries?

kai-fu lee: specifically, the impact of artificial intelligence on the economies of developing countries varies depending on their level of development and specific circumstances. take china as an example. although china is a developing country, china has been able to compare with the united states in terms of ai development and innovation. however, the situation is different for regions with relatively backward economic development, such as africa, south america and southeast asia.

these regions may face challenges because they may rely on imitating the development model of china or india. the chinese model relies on manufacturing, using low-cost labor to attract outsourced manufacturing operations to drive economic growth. the indian model, on the other hand, focuses on the outsourcing of white-collar jobs, taking on jobs that the western world is unwilling to do or that pay lower wages. however, the sustainability of both models will be challenged as robots and software threaten to replace these jobs in the future.

therefore, developing countries need to explore new development paths. these countries should consider developing all three skills mentioned above (i.e., creativity, empathy, and dexterity), but may have the greatest potential for expansion in developing empathy, for example by developing tourism or providing elderly care services . we have already seen some southeast asian countries begin to export healthcare service professionals to other countries. these jobs are unlikely to be replaced by artificial intelligence. at the same time, these countries should also try to cultivate a small group of creative talents as much as possible, because this will bring significant economic benefits, although resources may be limited and only a few people can achieve this. dexterity is also another area that can be developed, such as making exquisite arts and crafts, or making high-end products like swiss watches. these are often difficult for artificial intelligence to replicate, or people prefer hand-made products. overall, the impact of artificial intelligence on the economies of developing countries is multifaceted, and these countries need to formulate new development strategies to adapt to the future economy driven by this technology.

04 suggestions for future education and career planning: do what you are good at

smith: do you have any advice that would apply to students everywhere? if they were to choose college or another educational path in the future, what would you advise them to study?

kai-fu lee: first of all, i suggest reducing the emphasis on rote learning, because humans cannot surpass artificial intelligence at this point. secondly, innovation capabilities should be cultivated. third, improving soft skills of communication, collaboration, empathy, and building trust is critical.

at the same time, i recommend following your heart and passion, because you will be at your best when you do something you truly love. in an era where artificial intelligence and humans compete for jobs, we must do what we do best. also, think about emerging careers. just as the internet has spawned many new careers, artificial intelligence will create many job opportunities that have not yet existed, and we should stay alert. for example, data scientist is a hot job now, but it didn’t exist 20 years ago; uber driver didn’t exist ten years ago. therefore, be proactive and explore.

artificial intelligence has broad prospects for application in sustainable climate and environmental issues, and it can play an important role in data collection, problem analysis and solution provision. first, ai can predict and monitor the effects of climate change by analyzing climate models and environmental data. second, ai can help drive the energy transition, particularly in solar and battery storage technologies. as the cost of solar energy decreases significantly, and battery technology continues to advance, distributed solar and battery storage are expected to become the main source of energy in the future, which will help reduce dependence on fossil fuels and reduce energy costs and thus environmental impact. positive impact.

in addition, the application of artificial intelligence and automation technology in manufacturing can improve production efficiency and reduce costs, which also applies to the food and agriculture fields. for example, vertical farming and stem cell-based 3d printed meat can reduce reliance on traditional agriculture, reduce land use and water consumption, while improving food security and sustainability. the application of artificial intelligence in agriculture also includes precision planting, crop monitoring, soil management, and pest and disease prediction. these technologies help increase crop yields, reduce resource waste, and improve the overall sustainability of agriculture.

however, the development of artificial intelligence also brings new challenges, especially in terms of energy consumption and carbon emissions. with the widespread application of artificial intelligence technology, the energy demand of data centers continues to increase, which puts pressure on both the power grid and the environment. therefore, developing more efficient ai hardware and algorithms, as well as using renewable energy to support ai operations, are keys to achieving sustainable development of ai technology.

smith: it’s worth noting that people in the tech community tend to emphasize the positive aspects of artificial intelligence, while those outside the industry may be more concerned about its potential negative impacts. i admire your books "ai superpowers" and "ai 2041" for their balanced discussions of the opportunities and challenges of ai. ultimately, do you have an optimistic or pessimistic view of artificial intelligence?

kai-fu lee: i firmly believe that i am an optimistic supporter of artificial intelligence. at the same time, i also hope that i am a rational optimist, not only paying attention to its potential negative impacts, but also actively looking for solutions. my goal is to be a solution-seeking optimist rather than a simple, unrealistic optimist. (tencent news guest author/wuji)