Experts, execs stress responsible AI

No country, including the United States, can single-handedly lead the future development of artificial intelligence, and only global cooperation can ensure that AI will remain safe and beneficial for all, said field experts and company executives at the ongoing Boao Forum for Asia Annual Conference 2024 in Boao, Hainan province, on Wednesday.

Their remarks were an attempt to address "AI anxiety" among sections of the Chinese AI industry consequent to the launch of text-to-video model Sora by US firm Open-AI.

Zeng Yi, a researcher of the Institute of Automation, which is part of the Chinese Academy of Sciences, and a member of the United Nations High-level Advisory Body on AI, said: "The US will not lead the future AI development alone, nor will China. Unless we move forward hand in hand, no country will."

Zeng emphasized that AI is a foundational technology and an enabler.

"The world is vast enough to allow China, the US and many other countries to find their own opportunities in building infrastructure and driving applications.

"If the culture of China differs from that of the US, then please provide a profound understanding of how a different culture might envision the empowerment of technology in the future and how we can deeply integrate science and humanity to offer more possibilities for the world."

The more pressing issue, he said, is how the world can unite to truly develop AI that is safer, more reliable and more trustworthy.

Stuart Russell, a professor of computer science at UC Berkeley and a leading AI researcher, said that competition in the field is by and large healthy, but the idea of a zero-sum game is a "huge mistake".

"If AI can be created to be safe and beneficial, and more capable than human beings, then it can function as a source of unlimited wealth," he said. "In that sense, there is no need to compete for such technology."

Russell further said that a more important issue is to coexist healthily and safely with this technology, which is still an open question for the world.

Chinese tech companies such as iFlytek, Alibaba Group, Tencent Holdings, Baidu Inc, ByteDance and Huawei Technologies as well as thousands of startups are scrambling to develop AI large models. Many of them have gathered momentum over the past year.

As of October, China had developed 254 AI large models with a parameter of at least a billion tokens each, according to a report released by the Beijing Municipal Science & Technology Commission. Tokens are chunks of text that AI learns from, while a parameter is used for evaluating numeric data.

Liu Cong, vice-president of iFlytek and head of iFlytek's research institute, said the company's SparkDesk AI large model has had great improvements and is expected to catch up with ChatGPT-4 in Chinese abilities by the middle of this year. More importantly, it is trained on a totally independent computing power platform.

"But we have to note that only when the large model is applied can it become an important engine for new quality productive forces," he said.

"Facing the future, China must get a clear understanding of its gaps with those of international levels and, meanwhile, leverage its joint forces on resources to accelerate the systematic construction of AI."

Liu said that focus should be sharpened on the social risks and ethical challenges that the AI tech may pose, in order to ensure AI becomes a tool to build a better world.