From technical hot topics to industry deep cultivation: the second half of AI competition focuses on implementation and governance

Securities Daily Journalist Liu Zhao

At the Boao Forum for Asia Annual Conference 2026, AI is undoubtedly the hottest topic. Sub-forums centered around AI are taking place one after another, with rising enthusiasm. Unlike previous years where discussions focused more on technological breakthroughs and model iterations, this year the outside world is more concerned with how AI can truly enter the industrial front lines, moving from “appearing very strong” to “being effectively utilized,” and how to maintain safety, responsibility, and governance standards while accelerating implementation.

The 2026 “Government Work Report” proposed the creation of a new form of intelligent economy. It aims to deepen and expand “artificial intelligence +,” promote the accelerated deployment of a new generation of intelligent terminals and agents, drive the large-scale commercialization of AI in key industries, and cultivate new business models and patterns that are inherently intelligent. Observing the discussions at this year’s Boao Forum in line with this policy direction, it is clear that AI is moving from a technological race to deep industrial cultivation; its competitive focus is no longer solely on models and computing power, but also on scene embedding, organizational restructuring, and the synchronous follow-up of governance systems.

From Technological Competition to Scene Implementation

“As AI has reached this point, the industry’s main concern is no longer whether there are new models, but whether it can create real value,” said Jiang Xiaojuan, former deputy secretary-general of the State Council, director of the National Data Expert Advisory Committee, and honorary president of the China Industrial Economics Association. She noted that in the era of artificial intelligence and digital economy, the role of industries and enterprises in innovation has significantly increased, and the traditional linear innovation path—from scientific discovery to technological development, and then to industrial transformation—is being restructured. Industrial sectors are no longer just the end link of result transformation; they are increasingly involved in the discovery and research of frontier technologies.

Academician of the Chinese Academy of Engineering, Tsinghua University chair professor, and director of the Tsinghua University Intelligent Industry Research Institute Zhang Yaqin summarized the current AI development into three major trends: moving from generative AI to agent AI, from information intelligence to physical and biological intelligence, and from single technology to “AI+” fully empowering various industries.

This judgment was echoed in multiple discussions. Business representatives present generally believed that AI is gradually breaking through its positioning as an auxiliary tool and becoming an important force in reshaping business processes and industrial logic. Dai Pu, co-president of Roland Berger’s global management committee, cited his team’s survey results of 200 companies, stating that over 90% of companies are dissatisfied with their AI investment returns. The issue lies not in the technology itself, but in the fact that many companies remain at the stage of scattered pilot projects, localized optimizations, or even “just adding a chatbot,” without genuinely restructuring processes, reorganizing data, and reshaping organizational structures around AI. He believes that only by embedding AI throughout the entire enterprise process and establishing an application system based on proprietary data and system transformation can AI investments cross the “value chasm.”

Current AI implementation scenarios are rapidly emerging. “The long-standing issue in the education sector of achieving high quality, large scale, and personalization is making breakthroughs with the advancement of AI applications,” said Cheng Qun, vice president of Yuanli Technology Group and director of the Artificial Intelligence Research Institute, to Securities Daily. The collaborative architecture of edge-cloud, the evolution of intelligent terminals, and the continuous improvement of underlying capabilities such as communication and computing power provide support for the large-scale application of AI.

Many attendees predict that the focus of AI competition will increasingly shift from model parameters and general capabilities to industry understanding, scene transformation, and business closed loops. Those who can integrate AI into real production and life faster will be more likely to take the lead in the new round of industrial transformation.

Application and Governance Must Advance in Tandem

As AI accelerates its entry into the real world, the importance of governance becomes increasingly prominent. Interviewees generally believe that while AI can improve efficiency, optimize resource allocation, and enhance service accessibility, it must never run “wild” without responsibility boundaries and institutional constraints; application and governance must advance in tandem.

Xue Lan, dean of the Schwarzman College at Tsinghua University and director of the Tsinghua University Institute for International Governance of Artificial Intelligence, told Securities Daily that current technological breakthroughs are clearly ahead of application and institutional building, which means that the industry must recognize AI’s potential while also advancing ecological coordination and rule construction simultaneously.

Participants believe that the medical and health scenarios present AI governance issues more prominently. Wu Wenda, president of Tencent Health and head of Tencent Life Sciences Laboratory, bluntly stated that high-risk tasks cannot rely entirely on AI agents; the responsible party must be human and cannot shift responsibility by saying “AI made this judgment.” Chi Yongshuo, president of corporate affairs at LexisNexis and chairman of Elsevier, also pointed out that the health system is highly complex, and AI applications must raise efficiency while being cautious of its “shadow,” promoting knowledge sharing, evidence-based decision-making, and equitable accessibility, but the premise is to control adverse impacts within manageable limits. Li Tongyin, vice president of strategy and innovation and executive editor at Cell Press, believes that AI in the medical and health field should not only focus on whether the output is “decent” but also pay attention to whether the input data is high quality and trustworthy, and whether the output results undergo judgmental thinking and contextual verification. A decision error here could have far more serious consequences than in general consumer scenarios.

Zhang Yaqin proposed that AI-generated content needs to strengthen identification, and intelligent agents must be traceable to responsible parties. A significant portion of the existing rules in the legal system still applies, but timely adjustments must be made to fill institutional gaps regarding new technological forms. At the annual meeting, some attendees candidly told Securities Daily that issues such as data security, algorithmic bias, model transparency, energy consumption, and international collaboration will become key variables affecting the long-term healthy development of AI. In other words, AI competition is not just a contest of technology and business models but also a competition of governance capability, institutional supply capacity, and ecological coordination ability.

Currently, AI is rapidly moving away from “concept fever” and “showcase fever” into deeper waters, emphasizing effectiveness and responsibility. On one hand, intelligent agents, new terminals, industry-specific large models, and AI-native applications are continually emerging, pushing “artificial intelligence +” from point exploration to widespread application; on the other hand, discussions around responsibility identification, data governance, risk control, and rule construction are also noticeably heating up. For the industry, what truly determines how far AI can go is not only the speed of technological iteration but also the depth of implementation and maturity of governance. Only by finding a more robust balance between innovation and regulation can AI better become a new driving force for high-quality economic development.