Experts who attended the 2021 World Artificial Intelligence Conference called for the development of responsible and trustworthy AI, as the sector is set to transform countless aspects of everyday life.

According to the "Trustworthy AI White Paper" released by the China Academy of Information and Communications Technology during the event, major risks and challenges of AI include application hazards caused by insecure algorithms, nontransparent algorithmic models, data-derived decision bias, confirmation of liabilities and privacy leaks.

"China is pretty much in the same league with other developed countries in terms of AI advancements, so there is a genuine lack of an existing set of proven rules to regulate this market," said Xue Lan, a professor at Tsinghua University who is also director of China's national professional committee on next-generation AI governance.

"We need to develop technology and formulate governance rules at the same time, which could be the biggest challenge compounding the sector's development," Xue said.

Back in 2019, China issued principles for next-generation AI governance, pledging to develop responsible AI. The directive addressed eight tenets－harmony and friendliness, fairness and justice, inclusiveness and sharing, respect for privacy, security and controllability, shared responsibility, open cooperation and agile governance.

Xue said China has undergone two stages of AI governance, namely "responsive governance" involving taking reactive measures when problems occur, and "concentrated governance" that adopts proactive approaches to tackle issues comprehensively.

"Agile governance, by comparison, manages to strike a balance in promoting technological innovation and balancing potential risks," he said, adding this is the guiding path forward.

One critical objective is to develop "reliable" AI technology, which can be characterized by its stability, explainability, privacy protections and fairness, said Tao Dacheng, director of the JD Explore Academy, an in-house research unit of e-commerce company JD.

"The first and foremost task is to find suitable methods to quantify algorithms, models and systems in the above four aspects," Tao said.

The corporate world has a big role to play. Tao said the company will fully utilize its suite of retail, logistics, health and technology industrial chain strengths to achieve industrial breakthroughs and contribute to more fundamental research.

Ant Group is training its proprietary AI technologies to simulate hacker invasions in a bid to enhance risk-detection capabilities and thus boost its defensive resilience.

According to Ant's Vice-President and Chief AI Scientist Qi Yuan, the technology is now being applied to risk-control scenarios like anti-fraud issues and safeguarding secure transactions in multiple occasions on its mobile wallet Alipay.

Executives also appealed for proper supervision over blockchain, aiming to make the technology more "manageable".

"The manageability of blockchains should include aspects such as being reviewable, editable, authorizable, manageable and auditable," said Da Hongfei, CEO of Onchain, a Shanghai-based company using technology to empower both the private and public sectors.

"This attempt will realize consensus-based on-chain governance, which can shield socially harmful information, block or intercept illegal transactions and end services that do not meet regulatory standards," Da said.

The scale of the global AI industry reached $156.5 billion in 2020, a year-on-year increase of 12 percent, according to research firm IDC.That number stood at 303.1 billion yuan ($46.92 billion) in China, up 15 percent from the previous year, said the China Academy of Information and Communications Technology.