AIthropology Lab
ESEN

Economic anthropology

The Third Obligation: Accra and the Debt the West Refuses to Repay

The Accra summit turns resolution A/RES/80/250 into an institutional agenda. This note asks whether recognising historical harm also entails restitution and examines how that question reappears in the political economy of AI.

·Martín González Senosiain
An open, empty archival box lined with cloth on a dark wooden table, beside a paper-wrapped object and a small card: an image of restitution and heritage awaiting return.

Between 17 and 19 June 2026, Accra hosted the High-Level Consultative Conference on the Next Steps to the Landmark United Nations General Assembly Resolution A/RES/80/250. Convened by John Dramani Mahama as the African Union’s Champion on Reparations, it brought together delegations from more than eighty countries, alongside regional bodies, communities of African descent, academic institutions and civil-society organisations. Press coverage recorded that presence; the final document, for its part, is formulated in the name of the participants and includes the 123 member states that had voted in favour of the resolution in March. Its purpose was not to commemorate a declaration but to translate it into a working agenda.

The result was the Accra Next Steps Commitments on Reparatory Justice, available for download on the official conference site of Ghana’s Ministry of Foreign Affairs. It is a forty-six-paragraph document that proposes eighteen strategic pillars. It encompasses truth and formal apology, legal justice, compensation, cultural restitution, decolonisation, debt relief, reform of the international financial institutions, corporate accountability, health, gender and climate justice, education, memory and the right of return for the diaspora. It also creates three follow-up mechanisms: a Global High-Level Advisory Council on Reparatory Justice, a Global Panel of Experts on Cultural Heritage Restitution and a Global Legal Panel on Reparatory Justice.

The conference came with national commitments that are still limited but politically significant. According to information gathered by the Ghanaian Times, the Netherlands announced the return of some two thousand objects to Ghana; Germany expressed its willingness to repatriate goods from the Bono Traditional Area; Denmark reaffirmed its apology and its support for preserving Christiansborg Castle as a site of memory; and France signalled its readiness for a scientific commission aimed at establishing the historical truth. The coverage of the closing session nonetheless requires precision: these are announcements, expressions of willingness and initial commitments, not restitutions already carried out.

The backdrop is resolution A/RES/80/250, adopted by the United Nations General Assembly on 25 March 2026 by 123 votes in favour, 52 abstentions and 3 votes against. The text declares the trafficking of enslaved Africans and the racialised chattel enslavement of Africans to be “the gravest crime against humanity”, recognises that they constituted violations of jus cogens norms, and calls for a good-faith dialogue on formal apology, restitution, compensation, rehabilitation, satisfaction and guarantees of non-repetition. The official voting record confirms that Argentina, Israel and the United States voted against; the United Kingdom and the twenty-seven member states of the European Union abstained.

The resolution does not by itself create a legally enforceable mechanism or fix a reparatory sum. But neither is it an empty declaration. It places reparation within an international language of responsibility, affirms the persistence of the harms, and tasks the secretary-general with a report on states’ actions. That shift matters, because it transforms a claim often presented as moral memory into a sustained problem of cooperation, restitution and political responsibility.

Three noes and two ways of understanding responsibility

Of the three negative votes, the United States’ was the one that set out the most detailed and accessible doctrine. In its explanation of vote, Washington acknowledged the atrocity of the trafficking but rejected a legal right to reparation for acts that, on its reading, were not illegal under the international law of their time. It also objected that the term “gravest” established a hierarchy among crimes against humanity, and criticised the idea that historical wrongs could be used to redistribute present-day resources among populations temporally distant from the victims.

The European Union’s position was different in tone, though convergent in its legal consequences. Its statement before the General Assembly describes the Atlantic trafficking as a tragedy without parallel, recognises its lasting effects and affirms a collective responsibility to combat the inequalities that persist. But the EU rejected the superlative, questioned the historical references and held that retroactivity and reparation claims do not fit with the established principles of international law. The United Kingdom made a similar objection, resting on non-retroactivity and the principle of intertemporality.

Argentina and Israel joined the negative vote. Without an official, stable and verifiable explanation comparable to those of the United States, the European Union or the United Kingdom, it is best not to attribute a common doctrine to the two, nor to turn a diplomatic hypothesis into a fact. The firm datum is their opposition to the approved text.

Against these positions, the Accra framework starts from another premise. At the closing of the conference, Mahama insisted that history does not require inheriting personal guilt, but rather taking responsibility for its consequences. That difference is decisive. This is not a confrontation between memory and amnesia: it is a confrontation between two ways of translating the past into responsibility. One accepts historical and moral recognition but resists the idea that sustained reparatory obligations should follow from it. The other holds that, when a historical relationship has produced wealth for some parties and dispossession for others, recognising the harm obliges one to repair its effects.

What Mauss saw before the lawyers

In The Gift, Marcel Mauss described systems of exchange articulated by three obligations: to give, to receive and to reciprocate. The gift is not an innocent transfer of goods. It creates relationships, commitments, prestige, rivalry and, at times, conflict. Refusing to reciprocate is not simply keeping a thing: it alters the social relationship that the exchange has set in motion.

The analogy has a necessary limit. Mauss was not describing a relationship of absolute domination, and racialised enslavement denied precisely the status of subject to those who were captured, transported and turned into property. There is, therefore, no equivalence between gift and slavery. But that very difference makes the problem more visible. The trafficking was not a frustrated reciprocal exchange but an extraction that turned bodies, labour, reproduction, territory and future into sources of accumulation for other economies and institutions.

The Maussian question comes afterwards, not before: what happens when those who benefited from that relationship recognise the harm but deny that such recognition produces an obligation to reciprocate? Reparation does not consist in pretending that there was an original reciprocity between equal parties. It consists in refusing to allow violence and dispossession to fall outside any relationship of responsibility merely because they happened in another time.

Karl Polanyi helps to sharpen the point. In The Great Transformation, he called labour, land and money fictitious commodities: realities that the market treats as exchangeable goods even though they were not produced for sale. Racialised enslavement went further. It did not add a fourth fictitious commodity but instituted a regime that transformed people and their descendants into transferable, heritable and calculable property. The United Nations resolution puts it more precisely: it speaks of a “definitive rupture in world history” whose scale, duration, systemic nature, brutality and lasting consequences continue to structure everyone’s life through “racialised regimes of labour, property and capital”.

The Cauca Valley offers a concrete Latin American scene. In Puerto Tejada, Michael Taussig and Anna Rubbo documented how a community of formerly enslaved people that, after the abolition of 1851, had built a free peasantry on land of its own was displaced, with the arrival of the sugar mills in the 1950s, into wage labour. It is not a simple continuity of slavery, but it is a transformation of dispossession into labour subordination. Esclavitud y libertad en el valle del río Cauca allows one to follow that history without turning it into an abstract allegory.

Marshall Sahlins lets us call this relationship an extreme form of negative reciprocity: an appropriation oriented towards obtaining the maximum possible while giving the minimum. Alvin Gouldner supplies the other side of the problem, the norm of reciprocity that sustains the expectation that receiving generates some obligation to respond. The Atlantic trafficking institutionalised a relationship in which extraction was not followed by reciprocation but by new forms of subordination. Abolition formally ended legal property over persons, but it did not by itself liquidate the inequalities, the spoliations and the racial regimes that this economy had consolidated.

The trap of quantification

Faced with a historical debt of this magnitude, the contemporary reflex is to ask for a figure. How much is owed, who must pay, who receives, for how long and by what criteria. Economic compensation is not superfluous to this conversation. It is a legitimate and central demand of African, Caribbean and African-descendant movements. The Accra document itself expressly includes it among its eighteen pillars and calls for an approach that determines its modalities.

The problem begins when the figure is imagined as closure. A payment can be indispensable without thereby becoming absolution. Michael Taussig showed how the fetishism of the commodity can conceal the relationships of domination that make value possible. Carried into this debate, the risk is that an agreed sum might turn a history of violence, accumulated wealth and racial hierarchy into a settled file: paid, and therefore concluded.

Accra for now avoids that closure. It does not fix a final figure or a single mechanism of calculation. Instead, it proposes a broad architecture in which compensation coexists with cultural restitution, debt relief, institutional reform, access to health, climate reparation, education, memory, return and the participation of the affected communities. The point is not to displace the material towards the symbolic, but to recognise that no isolated compensation can by itself repair a historical relationship of dispossession.

This is why Accra is more interesting for what it sets in motion than for what it promises to resolve. The return of objects, archives and human remains can have concrete material, historical and affective effects. But its meaning does not depend only on how many pieces return or what figure is agreed. It depends on whether those gestures cease to be treated as voluntary concessions and become part of a sustained relationship of responsibility, restitution and institutional transformation.

The regime has not closed; it has also changed its raw material. If slavery turned people into property and Polanyi showed how the market treated labour, land and money as commodities, the economy of artificial intelligence opens another frontier: language, behaviour, relationships and data turned into input for calculation. Nick Couldry and Ulises A. Mejias call this process data colonialism: a form of appropriation of social life that turns the capture of data into a condition of accumulation. The continuity is not an identity between epochs, but a question about the operation that reappears: who turns the lived into a resource, who controls the infrastructure that processes it and who bears its costs. Models need data, compute, data centres, electricity, water and critical minerals, while infrastructure and computing capacity are concentrated in a few economies and companies, as documented by UNCTAD and the International Energy Agency.

There is a coincidence that is best read as an interpretation, not as a cause of the vote. The United States leads Pax Silica, a coalition geared towards securing the chains that feed AI, from critical minerals and energy to semiconductors, compute and infrastructure. Argentina signed the initiative’s declaration on 25 June and its government presented the country as a “reliable supplier” of critical minerals and strategic resources for AI. Alongside that positioning, the Argentine state has launched further initiatives: an AI Unit Applied to Security authorised to analyse historical data to predict crime, patrol networks and analyse camera feeds, and a Social Digital Twin, whose announcement opened public questions about data, surveillance and a possible relationship with Palantir solutions. Israel, also a member of Pax Silica, concentrates a growing share of its technology economy in AI: according to Startup Nation Central data reported by CTech, the funding of AI companies rose from 4.9 billion dollars in 2024 to a projected 7.9 billion in 2025. In Gaza, the press has documented that AI-based military systems made it possible to generate target lists at an unprecedented speed and scale, a claim that is best kept attributed to that research and not turned into a self-sufficient datum. The Washington Post reports, for example, that assessment by a specialist at Tel Aviv University.

But Pax Silica does not single out the three states that voted against. The coalition also includes the United Kingdom and, after the June enlargement, Germany, the Netherlands and the European Union: the first three abstained on the resolution; in the German and Dutch cases, their governments also made announcements or gestures of restitution in Accra. The rhyme is therefore structural, not a property of three governments. The Atlantic Council warns that the project still depends on external mineral processing: even a “reliable chains” strategy can keep the extraction of raw materials separate from the places where value, technology and decision-making are concentrated. None of this supports the claim that the United States, Israel and Argentina voted against A/RES/80/250 because of AI. Rather, it shows a shared geopolitics that treats minerals, energy, compute and data as resources to be secured. The Accra document itself already incorporates this tension: its pillar on technology calls for effective access to emerging technologies, including AI, and the removal of the barriers that prevent developing countries from benefiting from them. It does not ask for digital charity; it asks for capacity, infrastructure and a less subordinate participation in the technological order to come.

Mauss’s third obligation offers no automatic legal formula. It offers a more uncomfortable question. When a society recognises that it has received wealth, power and lasting advantages from a relationship founded on violence, can it still regard restitution as a generous option? Accra answers no. Reparation does not cancel history. It compels history to remain a relationship.

Sources and bibliography

Primary and documentary sources

AI, supply chains and data colonialism

Theoretical bibliography