<p>While AI-assisted decision-making is rapidly gaining global traction as a novel governance paradigm, the foundations of its legitimacy are yet to be fully revealed. To address this, we examine how AI-assisted decision-making affects public support and perceived decision quality, focusing on AI literacy. Using a survey experiment in China with a rural public goods provision scenario, we find two key effects. First, a digital legitimation effect where AI-assisted decision-making significantly increases public support compared to bureaucratic decision-making. Second, a digital compensation effect among individuals with low AI literacy, enhancing perceived scientific validity, internal political efficacy, and acceptance of AI-assisted governance at micro- and macro-levels. These findings offer a dual-pathways framework for legitimacy, showing that AI not only reinforces public support but also mitigates AI literacy disparities, fostering inclusiveness in governance decisions.</p>