AI in China
Over the past 15 years, Li Fei-Fei, director of the Stanford Artificial Intelligence Lab and the Stanford Vision Lab, has been working on algorithms that enable computers and robots to perceive and learn from the world. She cherishes the idea that machines will eventually understand the world and society in which humans live.
In early 2017, Li joined Google Cloud AI and Machine Learning, believing that as AI technology reaches more people, it will have a more profound industrial and academic impact. In an interview with NewsChina, she shared her unique views on the prospects of AI, its market potential, and challenges for this rapidly-growing field.
NewsChina: How will AI impact our lives?
Li Fei-Fei: AI is a science and technology with a history of more than six decades, and its main function is to make machines and computers more intelligent. During the latest boom in AI, data and algorithms brought forth a breakthrough in the computing capacities of AI. I believe that AI will impact our lives in many aspects but in most cases, AI won’t be a distinct entity but will provide greater convenience. For example, doctors needing assistance in diagnosis, the management of medical data and chronic diseases. AI will play an increasingly large role propelled by data and algorithms. In transportation, AI has altered travel through self-driving cars, computing optimization of big data and even the matching of supply and demand. I believe AI will impact various vertical industries, including agriculture, finance, healthcare and entertainment.
NC: How will AI be present?
LFF: Like the best technology, you won’t be able to feel it. Take electricity for example. Nowadays, we take electricity for granted as the power source of lights, elevators and computers because electricity has become part of our daily life and we are not likely to be consciously aware of its existence. It is the same situation for AI. It is not important whether we are aware of the existence of AI; what is crucial is that the innovative technology could make our society and basic necessities of life better and more convenient.
NC: Can AI eventually replace humans?
LFF: I think we should be very careful about using the word “replace.” Looking back to past scientific and technical advancements as well as the new industrial revolutions in the 20th century, you will find that to a great extent science and technology played a crucial role in facilitating our work. Technological advances made professional life easier for journalists, teachers, and doctors.
Nowadays it is not rare for machines to write news stories but it would never compete with journalists in investigative reporting and news analysis. The relationships between AI and humans are in most cases cooperative or supportive. Look at doctors. It is unlikely that AI will take over the job of doctors but what I want to see in the future is AI algorithms helping doctors offer better and more effective diagnosis and treatment for patients.
NC: Where do you expect AI will play a bigger role?
LFF: I expect to see that AI will play a bigger role in health care, making medical resources more accessible to the general public at a time when many people worldwide still lack basic health care. Google Brain has recently published its studies using AI to conduct medical diagnosis including diabetes checkups through scanning the eyeball, and diagnosis of breast cancer through AI. You can imagine that doctors have to read a lot of medical images and AI could play a helpful role in helping doctors make quicker and more accurate judgments.
In addition, AI could also play a significant role in long-distance medical checkups and diagnoses. For example, in impoverished and remote areas, many people are unable to access timely treatment and AI could manage preliminary diagnosis before securing the long distance medical resources for patients.
AI can also help to improve the management of chronic diseases. After being discharged from hospital, patients with chronic diseases need to take medicines to recover and AI could play a role in monitoring and managing the lives of patients. Aging is emerging as a problem in many countries, and it has become an urgent issue to take care of elderly people. AI offers a variety of choices to take care of seniors.
NC: In AI research, is there a difference between academia and industry?
LFF: In the process of R&D, both academia and industry have played a crucial role especially for the cutting-edge science and technology of AI. The main purpose of academia is to cultivate students. As a professor, my own work centers not only on scientific innovation but also on the delivery of knowledge, which is a crucial value of academia.
Another issue is that in academia, more attention is paid to fundamental research. There is great potential for the development of AI and sometimes it takes a short time for research findings to be developed into products, just like deep learning. Sometimes, however, it takes a much longer time for the transformation. For example, scientists spent more than 20 years researching the field before inventing the machine learning algorithms. It requires great patience and foresight to conduct fundamental research. In industry, however, product creation is the priority.
NC: How can academia and industry cooperate to boost AI development?
LFF: AI is a great driving force for the fourth industrial revolution. This is a great chance for partnership but it needs careful thought about the best ways to cooperate. I ventured into the AI industry through the Google platform because cloud computing has a great impact on many fields. What I want to do most is to apply AI to the fields that the general public need most rather than simply drawing up blueprints.
NC: How do you see the recent AI start-up boom?
LFF: It’s a chicken-and-egg situation; what is important for entrepreneurs is to pinpoint the products they want to make and find the solutions to the problem either through AI or other technical means. It is the only way to make products popular in the market. AI is only a tool and our goal is to use the tool to solve problems.