AI Empowers High-Quality Development of People's Livelihood—Overview of the AI for Better Life Thematic Report Series

On April 14, the AI For Better Life Forum of the 2026 World Internet Conference Asia-Pacific Summit was successfully held in Hong Kong. Government officials, representatives of international organizations and enterprises, and experts and scholars gathered to conduct in-depth discussions on key topics including the inclusive application of artificial intelligence, bridging the digital divide, smart government innovation, and global AI collaborative governance.

During the forum, eight outcomes under the World Internet Conference Think Tank Cooperation Program were released. Among them, three core reports—on the AI open-source ecosystem, AI evaluation system, and localized smart government services—outline a high-quality development pathway for artificial intelligence to serve people's livelihoods and empower society from three dimensions: technological inclusion, security and trustworthiness, and people-centered governance.

AI Open Source: From Technological Equity to Global Governance, Unleashing the Flywheel Effect for Inclusive Livelihoods

ModelScope, together with multiple industry, academic and research institutions, released the report "Global Value and Practical Exploration of the AI Open Source Ecosystem: Unleashing the Flywheel Effect and Forging a "Born Global" Consensus". Based on over 10,000 high-quality questionnaires and in-depth surveys of nearly 800 award-winning teams, the report reveals three transformative trends in the AI open-source ecosystem.

The report notes that AI open-source is evolving from code openness to resource inclusiveness. Technological "white-box" approaches break down black-box barriers, enabling ordinary entrepreneurs and cross-domain developers to access top-tier AI resources on an equal footing. Micro-innovators such as "one-person companies" have become the backbone of the ecosystem, contributing 95 percent of community applications, while 96 percent of developers are willing to open-source their outputs secondarily, forming a flywheel ecosystem of "access–application–feedback". China's AI sector has transitioned from a follower to a leader, with "born global" emerging as an industry consensus. With an open stance, China has deeply integrated into global collaboration and become a core contributor to the AI open-source ecosystem. Open source has become the optimal approach to enhancing AI security and advancing tech for good, delivering low-cost solutions for livelihood scenarios including visual impairment assistance, elderly care, and ecological protection. It channels technological dividends directly to communities, shifting public welfare from material distribution to capability empowerment.

AI Evaluation: Cognitive Alignment and Scenario Deepening, Building a Secure Foundation for Trustworthy and Inclusive AI

China Telecom Beijing Research Institute and China Telecom International Co., Ltd. jointly released the report "Cognitive Alignment · Scenario Deepening · Ecosystem Collaboration: The Core Paradigm and Path of AI Evaluation in the Future". The report states that AI evaluation is no longer a "supporting tool" for AI technology, but a core piece of infrastructure that determines the secure, controllable development and large-scale industrial deployment of AI. It will evolve around three trends: cognitive alignment, scenario deepening, and ecological collaboration.

The report puts forward three implementation pathways:

Centered on "cognitive alignment" principles, elevate AI evaluation from purely testing technical performance to assessing the essence of intelligence, accurately pinpointing the core of AI capabilities.

Adopt vertical scenario-based evaluation to align with real livelihood needs, drive in-depth adaptation of AI technologies across application scenarios, and support large-scale industrial implementation.

Rely on platform-based and governance-oriented collaboration to build a global trusted evaluation ecosystem, effectively addressing three practical challenges: technological adaptation, scenario deployment, and ecosystem construction, safeguarding the secure and orderly development of global artificial intelligence.

Smart Government: Localized AI and Human-Machine Collaboration, Building a Caring Governance Loop

Macao Association for Internet Research released the report "From Data Perception to Intelligent Decision-Making: Research on Constructing a "Smart Governance" Closed-Loop System Based on Localized AI". The report addresses three major pain points in current smart governance: cultural and contextual barriers, data silos, and concerns regarding data sovereignty and security.

It innovatively proposes a "social intelligence" solution integrating AITL (AI-in-the-Loop) and HITL (Human-in-the-Loop), building a dual-loop LECA (Perception – Interaction – Creation – Action) system. By developing localized knowledge bases and implementing classified data deployment, the system resolves the "acclimatization" issues of generic AI, shifting government services from passive response to proactive perception and data-driven delivery. Using Macao as a case study, the system significantly improves government response efficiency, Cantonese semantic accuracy, and risk prevention capabilities, forming a replicable and modular framework. It provides practical references for complex-scenario smart governance across the Greater Bay Area and globally, making government services both technologically advanced and people-centered.

The three reports complement and reinforce one another, jointly affirming that artificial intelligence can truly reach every household and serve social development only by upholding the core principles of inclusion, trustworthiness, people-orientation, and collaboration, breaking technical barriers through open-source, safeguarding security with scientific evaluation, and focusing on livelihood scenarios via people-centered governance, to co-create a smarter and better life for people worldwide.