The rapid diffusion of artificial intelligence (AI) is loosening the link between human labor and the value that firms create. Existing research has largely examined the distributional consequences of AI through labor-market supply and demand, factor prices, and changes in the aggregate labor share. This paper takes a different approach. It argues that value created through AI does not accrue to any party automatically; its allocation among wages, executive compensation, dividends, retained earnings, and investment is determined by distributional decision rights within the firm. Drawing on Yuji Ijiri's theory of accounting measurement, the paper further shows that the benefits and sacrifices of AI adoption may fall on different parties, and it distinguishes livelihood security from the distribution of additional value as two separate layers of the income problem. On this basis, it proposes three concepts: the AI Value Attribution Problem, the Separation of Benefit Recipients and Sacrifice Bearers, and the Two-Layer Structure of Livelihood Security and Additional Distribution. The paper concludes that wages in the age of AI should be reconsidered not merely as compensation for labor but as an institutional form of distribution through which value created by corporate activity is attributed to workers. It is an exploratory theoretical study that sets out an analytical framework ahead of empirical work.
