Artificial intelligence (AI) is increasingly embedded in the infrastructures that organise access to knowledge, economic opportunities, public services and social participation. However, understanding AI inequality requires more than examining whether algorithms produce biased or discriminatory outcomes. It also requires asking who owns the infrastructures supporting AI, who controls computational resources, who determines technological priorities and who benefits from the value generated by these systems. In The Inequality Regime of AI: Power, Allocation, and the Struggle for Justice, Massimo Ragnedda and Maria Laura Ruiu develop a sociological account of AI inequality that connects these questions of infrastructure, ownership, governance and redistribution. The book examines alternatives to concentrated technological control through two related concepts: Algorithmic Commons and Redistributive Infrastructures.
These concepts shift attention from correcting individual algorithmic outputs towards transforming the material, institutional and epistemic arrangements through which AI is developed and governed.
What are Algorithmic Commons?
Algorithmic Commons refer to arrangements through which data, computational resources, algorithms, knowledge and the infrastructures supporting AI are collectively governed and maintained, rather than being exclusively controlled by private actors.
The concept draws attention to the concentration of computational capacity and predictive power within proprietary technological infrastructures. Access to AI does not necessarily translate into the ability to influence how AI systems operate, whose interests they serve or how the benefits they generate are distributed.
In The Inequality Regime of AI, the algorithmic commons provide an alternative to the enclosure of digital resources. Rather than treating data and computational intelligence exclusively as assets for private accumulation, a commons-based approach considers how these resources might be governed through shared responsibilities, collective decision-making and accountability to the communities affected by their use.
This perspective builds on established scholarship on the governance of shared resources, particularly Elinor Ostrom’s (1990) work on collective institutional arrangements and David Bollier and Silke Helfrich’s (2019) account of contemporary commons.
However, the algorithmic commons should not be understood simply as open-source AI, unrestricted access to datasets or the public ownership of technological infrastructure. A resource does not become a commons merely because it is accessible. Its governance arrangements matter: who establishes the rules, who participates in decisions, whose interests are represented and how the resource is protected against enclosure or appropriation.
The distinction is important because an openly available AI model can still depend on concentrated computing capacity, privately controlled cloud services and unequal access to technical expertise. Openness at one level does not necessarily redistribute power across the entire technological infrastructure.
The book discusses several initiatives that illustrate alternative approaches to digital governance, including cooperative digital platforms such as CoopCycle, Barcelona’s data commons and Te Hiku Media’s Māori-language data initiatives in Aotearoa. These examples differ in their organisation and purposes, but they help illustrate how collective stewardship and community-defined priorities can challenge exclusive forms of technological control.
Algorithmic commons therefore raise a fundamental question: how can the resources supporting artificial intelligence be governed as shared social capacities rather than exclusively as instruments of private accumulation?
What are Redistributive Infrastructures?
Redistributive Infrastructures are technological, institutional and governance arrangements designed to redistribute not only material resources, but also epistemic authority, recognition, participation and care.
In Chapter 11 of The Inequality Regime of AI, Ragnedda and Ruiu argue that AI inequality cannot be adequately addressed through ethical guidelines, technical audits or corrective measures applied after technological systems have been developed. Such interventions may be necessary, but they leave unresolved the underlying distribution of resources, ownership and decision-making power.
Redistribution, in this framework, must become a principle informing the design, financing, ownership and governance of AI infrastructures.
The concept includes three interconnected dimensions.
1. Material redistribution
AI development depends on access to data, computational capacity, networks, technical expertise and other material resources. When these resources remain concentrated among a small number of corporations or institutions, other organisations and communities become dependent on infrastructures they cannot meaningfully govern.
Material redistribution therefore concerns the availability, ownership and governance of the resources necessary to develop and use AI. Public investment, shared computational resources and community-governed data infrastructures are among the possible approaches discussed in the book.
2. Epistemic redistribution
AI systems do not simply process information. Their development involves decisions about what counts as relevant data, legitimate knowledge and meaningful intelligence.
These decisions can privilege particular languages, cultural assumptions and forms of expertise while marginalising others. Redistributing epistemic authority means expanding the capacity of different communities to define technological problems, contribute their knowledge and influence the purposes for which AI systems are developed.
This dimension connects the argument to scholarship on Indigenous data sovereignty, including Kukutai and Taylor (2016), which examines the relationship between data governance, collective rights and self-determination.
3. Ecological redistribution
AI infrastructures depend on energy, water, minerals, land and global supply chains. Their environmental costs and benefits are not necessarily distributed equally.
Redistributive infrastructures therefore include questions of ecological responsibility: who benefits from computational development, who bears its environmental costs and how technological resources can be governed in ways that acknowledge ecological limits.
Together, these dimensions extend redistribution beyond the distribution of economic resources. They connect material capacity with the distribution of knowledge, decision-making authority, participation and responsibility.
How are Algorithmic Commons and Redistributive Infrastructures related?
Although closely connected, Algorithmic Commons and Redistributive Infrastructures are not interchangeable concepts.
Algorithmic Commons foreground the collective governance and stewardship of shared technological resources. They ask how data, knowledge, computing capacity and algorithmic systems can be organised beyond exclusive proprietary control.
Redistributive Infrastructures foreground the reorganisation of resources, authority, participation and responsibilities through technological infrastructure. They ask how the conditions supporting AI can be redesigned to produce more equitable distributions of power and social benefits.
Their relationship is therefore complementary. The algorithmic commons provide a way of thinking about collective governance, while redistributive infrastructures establish a broader framework for considering what must be redistributed and how governance should be organised.
This relationship is developed further through the concept of infrastructural democracy in The Inequality Regime of AI. Infrastructural democracy refers to embedding democratic governance in the design, maintenance and oversight of AI systems.
Such governance requires more than transparency. It entails meaningful participation from the definition of technological problems through implementation, evaluation and revision. It also requires recognising that affected communities should have the capacity to question technological interventions and, where appropriate, reject systems that conflict with their autonomy or collective values.
The central issue is not merely whether people can access AI, but whether they can participate meaningfully in shaping the infrastructures that increasingly organise their lives.
Why is correcting algorithmic bias insufficient?
Research on algorithmic bias has demonstrated the importance of examining discrimination, unfair classification and unequal outcomes. Yet a focus on individual systems can leave the structural organisation of AI largely unchanged.
A system might become more accurate or perform more equitably according to a particular fairness criterion while remaining dependent on infrastructures controlled by actors with little accountability to the people affected by its decisions.
Similarly, making a model or dataset publicly available does not automatically resolve inequalities in access to computing resources, technical expertise, institutional influence or the ability to challenge technological decisions.
Algorithmic Commons and Redistributive Infrastructures shift the analysis from individual outputs towards the arrangements that make those outputs possible.
This involves asking:
- Who owns and governs the data, models and computational resources supporting AI?
- Who determines which problems AI should address and which forms of knowledge inform its development?
- Who participates in technological decision-making, and can affected communities influence or reject proposed interventions?
- How are the economic benefits, environmental costs and social risks of AI distributed?
These questions connect AI governance to wider debates about social inequality, democratic participation, data sovereignty and the political economy of technological development.
What would more democratic and redistributive AI infrastructures look like?
Algorithmic Commons and Redistributive Infrastructures do not constitute a single institutional blueprint. Instead, they establish a framework for evaluating existing arrangements and considering possible alternatives.
Several principles follow from this framework.
First, collective governance must be supported by material capacity. Participation is limited when communities lack access to the computational resources, expertise and funding necessary to exercise meaningful control.
Second, participation must extend to the definition of technological purposes. Communities should not merely be consulted after important decisions have been made; their knowledge and priorities should inform the design and governance of the systems affecting them.
Third, redistribution requires institutional and legal support. Collective rights over data, public-interest investment and accountable governance arrangements can help protect shared resources against renewed concentration of control.
Fourth, alternative infrastructures must remain attentive to inequality within communities themselves. Collective governance does not automatically eliminate unequal influence, and participation may reproduce existing hierarchies when some groups possess greater authority or resources than others.
Finally, ecological responsibility must be incorporated into technological governance. The sustainability of AI cannot be evaluated solely through computational efficiency or economic returns. It also requires considering the material resources consumed and the distribution of environmental costs.
These principles suggest that the challenge is not simply to make existing AI systems fairer. It is to establish the conditions under which AI resources, knowledge and decision-making authority can be governed more collectively and distributed more equitably.
Conclusion: From technological access to collective control
Algorithmic Commons and Redistributive Infrastructures offer a way of connecting AI governance to broader questions of social justice.
The algorithmic commons challenge the exclusive control of shared technological resources, while redistributive infrastructures address the unequal distribution of material capacity, epistemic authority, participation and ecological responsibility.
Together, they shift the question of AI justice from whether technological systems produce fairer outputs towards how the infrastructures supporting AI are owned, governed and organised.
This perspective does not assume that collective governance is automatically equitable or that alternative infrastructures can be developed without political and economic constraints. Rather, it identifies the distribution of resources and decision-making power as central problems that technological innovation alone cannot resolve.
The contribution of The Inequality Regime of AI is to place these questions within a sociological framework connecting technological infrastructure with power, inequality and collective responsibility.
Ultimately, the question is not only whether AI can be made more efficient, transparent or accurate. It is who should control the infrastructures of artificial intelligence, whose knowledge should shape their development and how their benefits and costs should be distributed.
Algorithmic Commons and Redistributive Infrastructures provide a conceptual basis for examining these questions and for considering how AI might be organised around collective governance, social needs and more equitable forms of technological development.
Frequently Asked Questions
What are Algorithmic Commons?
Algorithmic Commons are arrangements for the collective governance and stewardship of data, computational resources, algorithms, knowledge and other resources supporting artificial intelligence. They offer an alternative to exclusive proprietary control and emphasise shared responsibility, accountability and collective decision-making.
What are Redistributive Infrastructures?
Redistributive Infrastructures are technological and institutional arrangements designed to distribute resources, epistemic authority, recognition, participation and care more equitably. In AI governance, this includes material resources such as data and computing capacity, as well as the authority to define technological priorities and responsibility for environmental consequences.
What is the difference between Algorithmic Commons and Redistributive Infrastructures?
Algorithmic Commons focus on the collective governance of shared technological resources. Redistributive Infrastructures focus on how technological systems can redistribute resources, authority, participation and responsibilities. The concepts complement each other but address distinct dimensions of AI governance.
Do Algorithmic Commons simply mean open-source AI?
No. Open-source software can contribute to shared technological development, but openness alone does not establish collective governance. The distribution of computing capacity, ownership, technical expertise and decision-making authority also matters.
How do these concepts relate to AI inequality?
Both concepts address the structural conditions under which AI inequality is produced. They shift attention beyond algorithmic bias and individual access towards ownership, infrastructure, governance and the distribution of technological resources and social benefits.
Where are Algorithmic Commons and Redistributive Infrastructures developed in the literature?
The concepts are developed in Massimo Ragnedda and Maria Laura Ruiu’s book The Inequality Regime of AI: Power, Allocation, and the Struggle for Justice, published by Routledge. Chapter 9 examines the algorithmic commons, while Chapter 11 develops the framework of redistributive infrastructures and infrastructural democracy.
Book and Further Reading
Primary source
Ragnedda, M., & Ruiu, M. L. (2027). The Inequality Regime of AI: Power, Allocation, and the Struggle for Justice. Routledge. https://doi.org/10.4324/9781003689379
Relevant chapters:
- Chapter 9, Digital Feudalism and the Algorithmic Commons, especially the discussion of commons-based infrastructures and infrastructural sovereignty, pp. 138–141.
- Chapter 11, From Inequality Regime to Redistributive Infrastructures, especially the discussion of redistribution, infrastructural democracy and collective governance, pp. 163–167.
Selected theoretical foundations
Bollier, D., & Helfrich, S. (2019). Free, Fair, and Alive: The Insurgent Power of the Commons. New Society Publishers.
Costanza-Chock, S. (2020). Design Justice: Community-Led Practices to Build the Worlds We Need. The MIT Press. https://doi.org/10.7551/mitpress/12255.001.0001
Kukutai, T., & Taylor, J. (Eds.). (2016). Indigenous Data Sovereignty: Toward an Agenda. ANU Press. https://doi.org/10.22459/CAEPR38.11.2016
Ostrom, E. (1990). Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge University Press.
Mazzucato, M. (2021). Mission Economy: A Moonshot Guide to Changing Capitalism. Penguin Books.