AI is often discussed through the language of ethical risk: discrimination, privacy violations, misinformation, opacity, unsafe systems and the displacement or intensification of labor. AI also raises broader questions of political economy and global justice. Contemporary systems depend on accumulated scientific knowledge, public investment, human labor, cultural production, data and infrastructures developed across societies. Yet the capacity to build, own and control the most advanced systems remains concentrated among a relatively small number of corporations based in technologically powerful countries.

This concentration creates an ethical challenge as significant as the behavior of individual systems and their algorithms. Who controls the infrastructures on which AI depends? Who captures the value generated from collectively produced knowledge and data? Which societies can shape the standards, priorities and development paths of this technology, and which remain primarily consumers of systems designed elsewhere?

The global technological divide extends far beyond access to commercial AI products. It spans the entire technological stack, from energy, minerals, semiconductors and data centers to cloud systems, datasets, models, standards and public expertise. Bridging it requires sustained investment in universities, research institutions, technical education, public computing capacity and locally relevant applications. It also demands greater support for models and datasets that reflect different languages, cultures and social priorities.

Because advanced infrastructure is costly and requires scale, few countries can achieve meaningful technological sovereignty through isolated national strategies. Regional and international cooperation will be essential. Technological sovereignty is best understood as a relational capacity: the ability of societies to make meaningful choices about the infrastructures on which they depend, align technological development with public priorities and cooperate internationally without surrendering policy autonomy.

President Xi Jinping recently called for a just and equitable system of global AI governance, emphasizing that AI should serve humanity, advance shared prosperity and help Global South countries strengthen their technological capacities. He also highlighted openness, international cooperation and broader access to the benefits of innovation. These principles offer an important basis for global dialogue at a moment when geopolitical competition risks narrowing the space for collective action.

China and the United States, as the two most influential centers of contemporary AI development, therefore carry a particular responsibility. Beyond technological performance, leadership in AI should be assessed through contributions to shared safety and human development, and support for the capacity of other countries to shape the technological future.

A further important step in this direction is to recognize AI's collective, social foundations. Advanced models are not created solely by the firms that commercialize them. They draw on generations of scientific discovery, publicly funded research, creative work, software development, data labor and everyday social interaction. This produces a fundamental imbalance between the socialized character of AI production and the concentrated character of its ownership and control.

Treating AI as a common good does not necessarily require rejecting private enterprise or imposing a single institutional model on every country. It requires recognizing that some foundational resources and capabilities have an inherently public and collective character. Computing capacity, research networks, selected datasets and general-purpose models could be supported through public, cooperative, regional or multilateral arrangements. Public institutions could use procurement, regulation and investment to reduce dependence on a small number of providers and orient innovation toward social needs. Such measures would preserve space for private initiative while ensuring that essential technological capacities remain subject to public purpose, broad access and oversight.

A common-good approach must also protect data workers and creative producers, preventing automation from eroding labor rights, and assessing AI's demands for energy, water, minerals and hardware. The greater good cannot be measured through technical performance or economic growth alone. It must encompass social inclusion, ecological sustainability and the fair distribution of benefits and risks. 

The author is an associate professor of sociology at University College Dublin, Ireland, and a part-time Associate Professor in the Department of Social Sciences at LUT University, Finland

Copyedited by G.P. Wilson

Comments to dingying@cicgamericas.com