Drawing upon the annual reports from the United Nations Conference on Trade and Development （UNCTAD） and the Stanford Institute for Human-Centered Artificial Intelligence （HAI），this paper systematically delineates three primary dimensions of artificial intelligence development in 2025. First，at the technological level，AI is undergoing a leap from quantitative to qualitative change. Frontier models are achieving breakthrough progress on numerous high-difficulty benchmarks，exhibiting performance convergence and fueling intensified market competition. Concurrently，AI applications are accelerating their penetration from the digital-virtual realm into the physical world，showing revolutionary potential and beginning large-scale commercial deployment in areas such as scientific discovery，precision medicine，intelligent manufacturing，and agriculture. Second，on the global landscape，a competitive dynamic led by the US-China duopoly is becoming increasingly entrenched. The United States，leveraging its significant capital advantages and deep research foundations，remains the leader in private investment and the development of top-tier models. China，meanwhile，leverages its massive data resources and state-driven strategic initiatives to lead in research output and patent volume，demonstrating astonishing speed in catching up on core model performance. The comprehensive strategic competition between the two nations in talent，hardware，and standards is profoundly shaping the trajectory of the global technology ecosystem. Finally，regarding its economic impact，the “double-edged sword” nature of AI is becoming increasingly evident. Empirical research broadly confirms that AI adoption significantly enhances firm productivity and exerts dual “equalizing” and “amplifying” effects on workforce skills. However，AI is also introducing profound structural shocks to the global labor market，posing a particular challenge to knowledge-intensive positions and raising widespread concerns about job displacement，rapid skill iteration，and social equity. This paper posits that artificial intelligence is at a critical historical inflection point. Its future trajectory is not determined by technology alone but depends on the collective choices made by global society concerning policy，regulation，and governance frameworks. In the face of the opportunities and challenges presented by this technology，establishing a human-centric，inclusive，and shared global governance system to steer AI toward a fair and sustainable path has become an urgent imperative.