In the context of the deep penetration of artificial intelligence (AI), how to address the malfunction of technological dividend distribution and achieve a synergistic coexistence of technological innovation efficiency and social equity has become a critical issue demanding urgent attention. This paper breaks away from the traditional assumption of neutral technology and constructs an integrated analytical framework of“technological shocks-market power reconstruction-institutional adaptation.” Drawing on the latest empirical data, it systematically analyzes China’s real-world dilemmas of employment polarization, income distribution imbalance, and insufficient institutional adaptation. The findings reveal that the routine-biased technological progress of AI triggers a polarization pattern in the labor market characterized by“expansion at both ends and contraction in the middle.”More critically, the advantages of algorithms and data reshape factor market power, reinforcing the algorithm-enabled buyer’s monopoly of firms, which leads to a pronounced phenomenon of wage markdown and widens the income gap between skill groups as well as between labor and capital. In response to the inadequacy of traditional institutions in adapting to new forms of production relations, this paper proposes the construction of an inclusive institutional system suitable for the AI era. While improving employment support and income distribution adjustments, efforts should focus on strengthening algorithmic governance and antitrust regulations, guiding technology through institutional innovation towards “empowering labor,”in order to achieve inclusive growth that balances efficiency and equity.