Researchers have developed a new algorithm that harnesses refugee data in order to better predict the outcome of placements in host countries.
A new study, developed by US and European researchers, claims that data algorithms could be used to much more efficiently place refugees in new communities and jobs.
The study, which focused on the US and Switzerland, used data from 36,000 historical refugee settlement cases to ascertain whether arriving refugees found employment. This information was then cross-referenced with biographical data of the refugee in question – age, gender, language and country and country of origin etc – to see if patterns could be developed to help predict successful placements.
According to a paper published in Science, the algorithm was able to reassign the 36,000 cases to more optimal locations, theoretically increasing refugee employment in the US from 25% to 50%, and in Switzerland by 73%.
Talking to The Verge, Jens Hainmueller, a professor in the Department of Political Science at Stanford University who worked on the study, explained this stark improvement is mostly due to the inefficient method currently being used to assign refugees. In Switzerland refugees are randomly assigned to one of the country's 26 administrative cantons, while in the US placement is purely based on capacity. Hainmueller adds:
“Right now, there’s no purposeful matching that happens at all. It’s literally just, if you come in a given week and there’s a spot in Denver — you’re going to Denver.”
Markus Spiske An algorithm to help refugees find a job is a step forward, but employment isn't the only factor when measuring success.
The system is not without its skeptics, however. Mike Mitchell, an overseer of refugee programs at the nonprofit HIAS, believes the algorithm ignores many other factors which impacts upon successful settlements, such as informal social connections and health issues.
Additionally, employment is not the only barometer of successful refugee placement. The political and social environment in host countries or individual communities also has an important impact, something currently unaddressed in the algorithm.
For their part, the study accepts these weaknesses, but suggest the algorithm can be used as a starting point for immigration personnel, with a 'human override' providing an extra layer of scrutiny.
The next step for the algorithm is to gain some experience in the real world. However, this could take a while, with Mitchell adding that inertia within immigration agencies and a hostile political climate regarding refugees could make any rapid adoption unlikely. For now, it seems like grassroots organisations are doing the most to help refugees find their way in their new homes, like the Brazilian online platform turning them into entrepreneurs and the UK social venture finding them work as language tutors when their foreign-earned qualifications aren't recognised.