Nikkei: China is on track to become an AI superpower, but has neglected the ethical limits of AI

Nikkei Asian Review reported that on June 25, the National Artificial Intelligence Security Commission, an independent U.S. government agency, released "The Role of Artificial Intelligence Technology in Responding to and Preparing for Pandemics: Recommended Investments and Initiatives." The report made 10 recommendations, including calling for the establishment of a federal "pandemic preparedness dataset" and investing in "digital modernization of state and local health infrastructure required for effective disease surveillance."

This is the third report on AI and the government response to the coronavirus pandemic released by the commission since May 6. The commission's so-called COVID-19 series of white papers addresses fundamental ethical issues related to the use of artificial intelligence in crisis response policies, including privacy and the responsibilities of software developers.

The commission seeks to help the administration develop clear positions on these issues and present them to Congress and the American people. This is time-consuming, but despite the urgency of the pandemic, the government must follow the democratic process when deciding how to spend taxpayer dollars.

In contrast, China has been very quick to use artificial intelligence to fight the disease. Artificial intelligence has been used for epidemic monitoring and prediction.

As of Thursday, the infection rate in the United States and Brazil was well over 15,000 per million people.

The pandemic has brought out huge differences in values ​​and cultural norms between countries. Shigeo Kawashima, an associate professor at Aoyama Gakuin University, believes that the world has reached a historic crossroads in terms of the evolution of artificial intelligence. The question, Kawashima said, is "which country will be able to build an AI-driven society in which people, not machines, play a central role."

Human interactions with technology are not culturally universal or value-neutral.

"Western society is based on Christianity and Judaism, while Chinese society is based on materialism and Marxism, so naturally they have different ethical systems," Kawashima said.

This is highlighted by an emerging research field called Explainable AI (XAI). XAI refers to a way of applying AI in a way that ensures that the decision-making process based on AI can be understood by humans. XAI can "explain" how it came to a specific decision.

The Defense Advanced Research Projects Agency (DARPA) of the United States Department of Defense is leading the way in research in this area. The U.S. military is keen to develop computer systems that can provide transparent explanations for the decisions made by artificial intelligence systems. XAI is a way to solve the "black box" problem in machine learning, in which even system designers cannot explain how their AI made a specific decision.

When computers see millions of images, they learn to tell the difference between cats and dogs. But using this approach, even if the AI ​​correctly identifies an image as a cat or a dog, people cannot tell how the AI ​​did it. Decisions based on machine learning are essentially a "black box." Artificial intelligence is an inherently opaque technology.

Different societies breed different kinds of AI. China’s AI is sure to get smarter at a dizzying pace, given the vast amounts of data available to its AI engineers.

According to a survey by the Japanese Patent Office in late July, Chinese entities filed 6,858 AI patent applications in 2017, surpassing the 5,954 applications filed by the United States. The United States has lost its long-standing overwhelming advantage in the field of AI. With patent applications doubling every year, China is on track to become an AI superpower.

Most AI patent applications made in China involve practical machine learning, and few are related to XAI. China may be the first country in the world to develop black-box AI technology in health care, autonomous driving, criminal trials, personnel decisions and other fields. But China's approach to AI research and development is completely divorced from ethical considerations, which inevitably raises concerns.

In April 2019, the European Commission's High-Level Expert Group on Artificial Intelligence released the "Ethical Guidelines for Trustworthy AI", which established seven key principles including transparency, diversity, non-discrimination, fairness and accountability.

In Japan and the United States, a large number of large companies including Google, Sony and Fujitsu have established their own AI ethics standards because without user trust, AI business will not survive.

Hirokazu Anai, senior director of Fujitsu Laboratories’ AI Lab, once said, “We have to prove to our customers that our AI is sufficiently safe and trustworthy.”