This keynote examines the landscape of AI innovation beyond the dominant axis, with particular attention to grassroots efforts in Global Majority countries focused on cultural preservation, linguistic expansion, and context-specific AI applications. These initiatives, many operating with minimal institutional resources and under pressing socioeconomic conditions, represent not a deficit of capacity but an instructive model of what it means to develop technology for communities rather than at them.
Against this moment of profound institutional disruption, this talk makes an affirmative case for a global public knowledge commons as both an epistemic and ethical imperative for advancing responsible AI development and broadening scientific inquiry. Such a commons would redistribute the benefits of AI while expanding capacity to shape its direction. In tandem with grassroots AI, a global public knowledge commons offers an empirical counter-argument to the ideology of AI accelerationism: that unrestricted, large-scale development is neither the only nor the most beneficial path forward.
