Planetary Ethics
The AI revolution is planetary in its reach, in its infrastructure, and in its consequences. The training corpus draws from every accessible human tradition. The data centers span continents. The affected populations include everyone whose labor was absorbed, whose data was harvested, whose cognitive environment is shaped by model outputs, or whose livelihood is reorganized by AI-mediated markets. An ethics adequate to this scale cannot be national, sectoral, or parochial. It must be planetary. But a planetary ethics that merely universalizes the values of the metropolitan center repeats the colonial move of treating particular values as universal. Mbembe's proposal: an ethics that holds both dimensions — the genuine interconnectedness that requires planetary-scale thinking, and the genuine inequality that requires attention to the specific conditions of specific populations.
The Infrastructure Lock — Contrarian ^ Opus
There is a parallel reading that begins from the material substrate of AI systems — the fiber optic cables, data centers, GPU farms, and energy grids that make planetary-scale computation possible. This infrastructure is not neutral terrain awaiting ethical governance; it is already owned, already located, already dependent on specific supply chains that run through specific territories controlled by specific states and corporations. The planetary ethics Mbembe envisions requires negotiation across traditions, but the negotiating table itself sits in a data center in Virginia, runs on chips fabricated in Taiwan, and depends on rare earth minerals extracted from the Congo. The infrastructure determines who can meaningfully participate before any ethical framework is proposed.
The temporal mismatch between infrastructure deployment and ethical deliberation is not merely unfortunate timing — it is the mechanism by which metropolitan control reproduces itself. By the time Lagos has a seat at the table, the technical standards are set, the platforms are built, the network effects have concentrated, and the ethical conversation is about mitigation rather than architecture. The communities whose data was harvested to train GPT-4 were not consulted about whether they wanted their linguistic patterns incorporated into a system they cannot access, run, or modify. Mbembe's framework assumes these communities could gain meaningful sovereignty over 'their' data, but the data has already been absorbed into model weights that are corporate property, running on infrastructure they do not control. The planetary ethics project, however necessary, arrives after the planetary infrastructure has already been built to concentrate rather than distribute power.
In the AI Story
The tension in the proposal is real and Mbembe does not try to resolve it neatly. Most planetary-ethics proposals in the AI space — from UNESCO declarations to the EU AI Act to the various industry self-governance frameworks — operate from the metropolitan center and distribute obligations outward in ways that preserve metropolitan authority over the framework itself. The developer in Lagos is governed by the framework; she was not consulted in its drafting. This is not planetary ethics; it is metropolitan ethics with global reach.
What Mbembe proposes instead is an ethics that requires substantive participation from the populations its frameworks would govern. This has procedural implications (governance structures must include meaningful representation from affected populations), substantive implications (the values encoded in the framework must be negotiated across traditions rather than imposed from a single tradition), and institutional implications (the bodies that develop and enforce the frameworks must be genuinely multi-polar, not merely formally so).
For AI specifically, the planetary-ethics framework generates several specific commitments. Data sovereignty for the communities whose data is used. Epistemic decolonization of training corpora. Participation of content moderation workers in the decisions that govern their labor. Redistribution of the value created by AI toward the populations whose work made it possible. Technical and institutional infrastructure that is not concentrated in a small number of metropolitan corporations.
None of this is straightforward, and Mbembe is too honest a thinker to pretend otherwise. The coordination problem is immense: getting many parties to the table, each with distinct interests and distinct capacities, and producing frameworks that none of them could impose unilaterally. The institutional problem is immense: existing international bodies (UN agencies, trade organizations, standards bodies) are themselves structured by the very hierarchies the ethics is meant to address. The temporal problem is immense: AI is developing on a timeline that outpaces the slow work of building legitimate governance.
What the framework provides is not a blueprint but an orientation. It tells us what success would look like and what specific failure modes to resist. It makes visible the difference between planetary ethics in substance and planetary ethics as metropolitan ethics in planetary drag. That clarity is itself a contribution, and it is the contribution Mbembe's work is best positioned to make.
Origin
The planetary-ethics framing develops through Mbembe's later work, particularly Out of the Dark Night (2021), in conversation with thinkers including Dipesh Chakrabarty on the planetary condition, Gayatri Spivak on planetarity, and the broader tradition of decolonial political theory.
Key Ideas
Planetary scale, not metropolitan universalism. A real planetary ethics must be negotiated across traditions, not extended from one tradition.
Participation is substantive, not procedural. Including affected populations means they have real authority, not advisory roles.
Multiple institutional implications. Data sovereignty, epistemic decolonization, worker voice, value redistribution, and multi-polar infrastructure are specific commitments.
Coordination is hard. The framework does not pretend the coordination problem is solvable by will alone.
Orientation, not blueprint. The contribution is to make visible the difference between real and performed planetary ethics.
Appears in the Orange Pill Cycle
Negotiable Sovereignty Zones — Arbitrator ^ Opus
The weight of truth shifts depending on which layer of the AI stack we examine. At the infrastructure layer, the contrarian view dominates (80/20) — the physical geography of computation is already set, with GPU production, cloud infrastructure, and submarine cables creating hard constraints on who can meaningfully participate. The infrastructure lock is real and shapes every conversation that follows. But at the governance layer, Mbembe's framing gains strength (60/40) — international bodies, however imperfect, do sometimes respond to sustained pressure from the periphery, as seen in debates over data localization, content moderation standards, and algorithmic accountability.
The question of timing reveals the sharpest tension. If we ask "what is possible given current trajectories?" the contrarian reading is almost entirely correct (90/10) — the speed of AI development and deployment far outpaces the slow work of building legitimate multi-polar governance. But if we ask "what orientation should guide our interventions?" Mbembe's framework becomes essential (70/30) — without a clear conception of what planetary ethics would actually mean, we cannot even identify which compromises are tactical retreats versus fundamental capitulations.
The synthetic frame might be "negotiable sovereignty zones" — spaces where the infrastructure lock loosens enough for genuine negotiation to occur. These zones are not given by the architecture; they must be carved out through technical standards battles, regulatory arbitrage, infrastructure investment, and the patient work of building alternative technical capacity. Mbembe's planetary ethics provides the orientation; the contrarian's infrastructure analysis shows where the leverage points might be. The framework's value is not in resolving the tension but in maintaining it productively — insisting on planetary ethics while acknowledging planetary physics.
Further reading
- Achille Mbembe, Out of the Dark Night: Essays on Decolonization (Columbia University Press, 2021)
- Dipesh Chakrabarty, The Climate of History in a Planetary Age (University of Chicago Press, 2021)
- Gayatri Chakravorty Spivak, Death of a Discipline (Columbia University Press, 2003)
- Boaventura de Sousa Santos, Epistemologies of the South (Routledge, 2014)
- Felwine Sarr and Bénédicte Savoy, The Restitution of African Cultural Heritage (2018)