Interview: "The ultimate power is love, not AI": top artificial intelligence expert

Source: Xinhua| 2017-05-31 23:16:47|Editor: Mu Xuequan
Video PlayerClose

by Xinhua writer Guo Shuang

LOS ANGELES, May 31 (Xinhua) -- "The ultimate power, in my opinion, is love, not AI," a top artificial intelligence (AI) expert and chief scientist of AI/ML (Google Cloud) Fei-Fei Li told Xinhua in a recent exclusive interview.

In the Chinese language, the abbreviation of artificial intelligence, or AI, is pronounced the same as the character for "love."

From the introduction of a second generation of its TensorFlow Processing Unit (TPU) to TensorFlow Research Cloud, almost all of Google's latest big announcements touched on AI in one way or another.

Last year, the technology giant recruited Li, one of the most eminent women working on AI, to lead machine learning research for Google's cloud computing division.

"Goal is to democratize AI for everyone!" Li, also director of the AI lab at Stanford University's computer science department and the Stanford Vision Lab, tweeted in January on her first day at Google Cloud.

"AI is the driving force of the fourth industrial revolution," Li told Xinhua. She believed that all vertical industries have the potential to benefit from AI deployments.

"From health care, transportation, retail and consuming, financial service, media and entertainment, manufacturing, agriculture, government, energy and oil ... all will benefit from AI," she said.

"AI can make life better: it can help doctors diagnose, deliver health care to remote areas, improve the environment, discover new materials, make energy more efficient ... I hope AI is a more welcoming field to people with diverse backgrounds," Li said.

What is the key to democratizing AI? Does it mean lowering the barriers? Li answered, "to reach as many people as possible."

"When you say lowering the barriers, it depends on who are you lowering them for. Of course that is part of it, but the end goal is to reach as many people as possible by using those AI-driven technologies," Li said.

"For example, if you are a consumer, how do we use AI products to help you have a better experience? You might not know if it is AI, or which algorithm is behind this, but it doesn't matter. We know how we should deliver it to you, so that is the end goal," she said.

"On the way to the end goal, lowering the barriers is very, very important, because when you do that, you can get more developers, more companies and more people to participate in the process of designing products, interfaces and business ... But the end is to reach as many people as possible," she said.

While opinions differ among AI experts, Li believed that humanity as a whole has the responsibility to use AI technologies to benefit mankind.

With more tools and technologies being developed, "we see how they could be used to benefit people or be used in pretty bad ways and cause big problems," Li said.

"It's not only responsibilities for developers or discoverers. I think all these people should be engaged: Silicon Valley leaders, professors, students, policymakers, lawmakers, educators ... everybody should be at the table discussing this."

Meanwhile, Li agreed that AI currently remains an extremely new and complex field, with resources in the area concentrated in a few leading companies and top universities.

Talking about the lack of women participating in AI research, Li said her effort is like "a drop in an ocean," adding that the tech and AI culture are still dominated by men.

"I think that needs to be changed. We need to encourage women and underrepresented minorities to get involved," she said.

"I want Silicon Valley leaders to talk more about this. If they all can go up and say that there should be 30 percent of women or something on AI research panels, that would be awesome," Li added.

As to the question of where we are now in AI development, Li said, "We are harvesting some of the early successes of deep learning, and we are also planting the seed for more innovative work to come."

Talking about her January trip to China, Li said she was there to learn. "I feel very positive for China. From the government to academia and industry, from entrepreneurs to technologists ... almost everybody is passionate about AI."

"I think that's wonderful, and that energy can transfer to a lot of innovation," she said.

The expert also said data is one of China's strengths. "China is a big market with a huge population, so there is a lot of data in China that can be used."

Li also expressed the hope that AI will become a bridge between the United States and China by providing more cooperation opportunities.

"There are shared needs and passion for AI ...I see that as a win-win," Li said.

Despite distractions from contemporary events, she said AI technologists and researchers need to remain focused on their task. "We still have to be grounded and cool-headed and to focus on developing good research and technology," she said.

"Sexy as AI is," Li said, "yet at the end of the day, it is deep mathematical engineering technology."