Some refugees have the employability and language skills to succeed anywhere. Some locations have strong labour markets and charitable companies that would benefit any refugee. These commonplace observations attract brilliant minds working in artificial intelligence and human geography at Oxford and Stanford. Can an algorithm solve the problem of burden-sharing in the international refugee regime? Does the destination matter? The data reveal clear synergies between individuals’ characteristics and local conditions. This suggests that some refugees’ strengths will be rewarded more in certain places than in others, while traits that might be liabilities in some places become less detrimental in others.

Countless refugees have disappeared after learning about the decision on their destination country. Others have vanished after their relocation, and preventing such irregular secondary movements has become a central policy goal. Member states of the European Union (EU) floated the idea that countries could make a financial contribution instead of being obliged to receive refugees from overstretched Border States like Greece and Italy. This notion met with Darwinian natural opposition so an alternative was offered that sought to distribute persons among member states on the basis of a distribution key. The Commission invoked Article 78(3) of the Treaty on the Functioning of the EU that triggered a temporary European relocation scheme for asylum seekers.

The distribution key was based on impartial, quantifiable and verifiable criteria that reflected the capacity of the Member States to absorb and integrate refugees, with appropriate weighting factors reflecting the relative importance of such criteria based on the following elements: (a) the size of the population (40% weighting) as it reflects the capacity to absorb a certain number of refugees; (b) the total of the GDP (40% weighting) as it reflects the absolute wealth of a country and is thus indicative of the capacity of an economy to absorb and integrate refugees; (c) the average number of unprompted asylum applications and the number of resettled refugees per 1 million inhabitants over the period 2010-2014 (10%) as it reflects the efforts made by Member States in the recent past; (d) the unemployment rate (10% weighting) as an indicator reflecting the capacity to integrate refugees, with a 30% cap of the population and GDP effect on the key, to avoid disproportionate effects of that criterion on the overall distribution. Finally, the actual numbers of refugees to be relocated to each member state depending on the total number of persons to be relocated.

Algorithms have taken full charge of politics in Brussels. Arbitrariness is dead. But algorithms alone can’t untie the knots around refugee relocation. They have however jump-started new work in political economy and have given a new face to the fading area of political policy that remained constrained for decades. Algorithms can show how refugees may contribute to competitiveness. But algorithms can also treat humans with hostility. Although the Council appealed to member states to give priority to particularly vulnerable persons including unaccompanied minors, pregnant women, disabled and elderly persons, some member states remained reluctant to receive persons from these groups.

Furthermore, although legally required to accept all types of refugees, some member states rejected allocations on the grounds that their preferences were violated. Algorithms remain tools that can validate the value that refugees bring to communities and can function to improve the decision-making process. But ultimately this “god from a machine” can only inform such decisions- not make them. They can only generate recommendations to case-officers who make the final determinations. This human override considers what an algorithm cannot take into account. The algorithm created by Stanford’s Immigration Policy Lab can identify the best location to resettle a refugee with an improved chance of being gainfully employed by as much as 70 per cent. It was trained on historical data of resettlements and was built by analysing refugee biographical information including: age, gender, language skills, country of origin, where they were resettled and whether or not they got a job.

For each country, the algorithm searches data to unravel what circumstances led to a refugee getting a job quickly and what circumstances may have resulted in the opposite effect. Using these patterns, the algorithm makes its own predictions about where an individual should be sent, and how well they may settle there. The Immigration Policy Lab noticed however that the algorithm works well with group data but collapses with individual cases because it cannot take into account countless constraining factors that enmesh what it is to be a human.

Escaping favelas, arrivals have few resources and must rapidly accommodate the receiving cultural melee. They remain economically marginalized for many years. West Indian spaces like Curaçao are at an interval in their plantation political economy that force them to explore fresh understandings that can underpin policy decisions as they strategize to make the resettlement of refugees from the Bolivarian Republic fairer and easier. Alexander Betts at Brasenose College, Oxford notes that around the Western world, politics has been convulsed by mutual vilification over migration. But for Betts, policy must always precede Acts and Algorithms. Many refugees are fleeing re-education camps, ethnic cleansing and political persecution and they are invaluable resources, not terrorists.