The logo of Chinese AI DeepSeek shown on a handphone on April 24, 2026 (CFP Photo)
Themed "Intelligent Partners, Co-Creating the Future," the upcoming four-day 2026 World AI Conference (WAIC) and High-Level Meeting on Global AI Governance will be held in Shanghai from July 17. The event will bring together over 1,400 guests from China and around the world, along with representatives from more than 1,100 enterprises.
Committed to providing international public goods, China proposed the Global AI Governance Initiative in October 2023. It facilitates and takes actions to foster an AI for good and inclusive development while supporting Global South countries in enhancing their AI capacity-building.
In a recent exclusive interview, 2007 Turing Award recipient Joseph Sifakis praised China's pragmatic approach to developing its AI strategy. He also envisioned an inclusive and sustainable AI-driven economy that promotes human well-being, supported by more ad hoc and concerted regulations for global AI governance.
A different AI prioritizing human well-being
No one doubts that AI is one of the key growth engines of the 21st-century global economy. However, economists and AI experts differ in their views on the extent to which AI can boost economic growth in advanced and developing countries.
According to the World Bank's latest Global Economic Prospects report, published this June, AI's greatest contribution is expected to come from faster productivity growth over the medium term. Under optimistic assumptions about AI’s productivity gains, economic growth in the 2030s could exceed the average growth recorded in the 2000s.
However, the gains would be considerably smaller under less optimistic assumptions. The 2024 Nobel winner Daron Acemoglu estimates that AI will increase U.S. GDP growth by 1.1% to 1.6% over the next decade, with an annual productivity gain of approximately 0.05%.
"To ensure an inclusive and sustainable AI-driven economy, we need a different kind of AI that prioritizes human well-being," Sifakis underscored.
Today, big tech giants are obsessed with the idea of Artificial General Intelligence (AGI) to build machines that are surpass human intelligence and outperform humans in any particular task.
Sifakis describes this as the American vision, which frames machines and humans as competitors, implying the supremacy of machines while, to some extent, seeking to marginalize humans. "They don't care about unemployment or ethical issues. All they care about is profit and market share," he said.
The Turing awardee doesn't share this vision. Instead, he warns that people should be aware of these problems, noting that current conditions are more favorable than ever for raising public awareness.
He cited issues of mistrust and resistance, especially in the US, stemming from the intrusive and uncontrolled application of AI. In addition, U.S. companies face challenges in generating sufficient revenue from the type of AI solutions they currently offer. As a result, their long-term economic viability remains uncertain.
Sifakis envisions a different approach known as Collaborative AI, in which machines provide an initial imperfect solution that humans can review, refine, accept, or reject.
However, he acknowledges that the working mechanism remains largely a black box requiring further research. He encourages countries to invest in this direction, saying, "We don't understand how they work. The quality of collaboration with humans is not very good. If we want to build Collaborative AI, we have to completely change the vision of AGI."
"AI is very important to the future of humanity to be left solely in the hands of companies and individuals who prioritize power and profit over the common good. It is also our responsibility to take action and try to change things," Sifakis concluded.
Joseph Sifakis delivers his speech online at the 2026 Forbes China AI Tech Summit in Shenzhen on July 10.
An edge for China to develop reliable autonomous systems
Speaking about the current status quo of AI development in China, the United States, and the EU, Sifakis—a member of the Chinese Academy of Sciences, the American Academy of Sciences, and Academia Europea—believes that China is well positioned to become a key player in the race toward autonomous systems.
"China can leverage its strong industrial base. We have many sectors that are demanding smart products and services. In addition, China can capitalize on the vast amount of data generated by its extensive and diverse digital infrastructure," he noted. According to Stanford University’s AI Index 2026, China had about 449 data centers in 2025, ranking second after the US.
What Sifakis admires about China is that it decides to build something, it invests whatever resources are needed. It also pursues that goal with unwavering determination, even in the face of potential failure. China can mobilize all the forces, as the state is strong enough.
In late June, the Turing laureate was formally appointed as chief scientific advisor to the AI innovation center of a major Chinese mobility platform.
Sifakis explained that the main reason for collaborating with the company is its close alignment with his research on autonomous transportation systems. They share a common vision for the future of robot taxis—it will not be determined by who has the largest AI model but by who can build the most reliable autonomous systems.
He believes that the future of autonomous systems will be defined by the successful integration of systems, engineering and algorithms. He also finds it encouraging to see the company making steady and tangible progress in this direction.
Moreover, as a distinguished visiting professor at Southern University of Science and Technology in Shenzhen since 2019, Sifakis has made regular visits to the Guangdong-Hong Kong-Macao Greater Bay Area (GBA) to deliver lectures and participate in academic conferences on autonomous systems from time to time.
He proposes that leading AI companies in the GBA should develop their own vision by capitalizing on their own strengths, namely, the very powerful industrial base.
According to Sifakis, their competence lies in sectors such as transportation, energy, telecommunications and smart cities. Rather than simply following the lead of big tech giants, they should develop their own broader vision for applying AI to industrial applications.
Sifakis observed that Chinese companies excel at integrative innovation. However, developing more groundbreaking, original ideas remains a key challenge for further progress. He also encouraged more integration between academic research and the rest of the economy in China through the case of Silicon Valley's ecosystems integrating industries and universities.
Long walk to AI global governance
Amid the rapidly evolving AI era, which gives rise to scientific, economic, social and industrial challenges, the Turing Award laureate urges all countries to reach a global agreement on how to manage AI-related risks.
There is no consensus on what can or should be regulated. It is essential to have effective mechanisms in place to enforce the application of these rules.
Sifakis illustrates the divergence between the EU and the US regarding AI regulations. The EU has comprehensive AI regulations, such as the EU AI Act, enacted in August 2024, and the Digital Services Act, enacted in August 2023. This means that the EU applies a risk-based approach to AI products, requiring higher reliability. In contrast, the US regulations are much less restrictive, and the country accepts the idea of self-certification.
"There are significant differences in how we approach AI-related risks, and the chances for a global agreement on AI are slim. Even the United Nations is striving to achieve this goal," he lamented.
Notably, the newly established World Data Organization in Beijing this March and China's efforts to launch the World Artificial Intelligence Cooperation Organization, currently scheduled to be headquartered in Shanghai, offer a silver lining for global AI governance.
"China is taking a pragmatic approach by implementing regulations tailored to specific application domains. I'm sure that China also applies technical rules to ensure the trustworthiness of AI and to guarantee that it is used responsibly," he noted.
Reporter | Zhang Ruijun
Photo | Zhang Ruijun
