He does not remember the nationality of that girl, but he does remember that she was eight years old, that she was traveling alone from the coast of Libya and that thanks to artificial intelligence (AI) she was able to find relatives in Europe. Mauro Di Si, search and rescue coordinator at the Open Arms organization, recalls this story in a call with EL PAÍS. “The child spoke little French. Through AI we were able to communicate with her and understand that she had family outside of Africa,” says Di Si. For the last two humanitarian missions in the Mediterranean, the Polish company VA Translators provided its V4 model to the organization. A device that translates voice, text or photographs instantly and detects 108 different languages, including regional languages of the African continent.
A small infirmary, but fully equipped as if it had nothing to envy of an advanced assistance ambulance, was located on the deck of the Open Arms rescue ship during the months of December 2023 and January 2024. Like a stethoscope or a first aid kit , these intelligent translators became another essential element for the doctors on board the NGO ship. “The devices have helped a lot when carrying out medical evaluations because with more agile communication, migrants establish a framework of trust with the health workers in which they can refer to traumatic scenes and thus, the doctors investigate to find a more accurate diagnosis. precise”, justifies Di Si. VA Translators, based in Barcelona, which has already used previous models of intelligent translation with Ukrainian refugees, donated four of these devices in November 2023 for the organization’s rescue missions.
In addition to health personnel, the cultural mediator assigned to this mission, who speaks five languages, has also used AI translators. “They allowed him to understand the most problematic situations in very particular cases of vulnerable people,” says Di Si. The rest of the crew, who speak English, also found it helpful to use them to communicate fluently with the rescued people. “That any worker or volunteer could talk to the migrants, without depending on the cultural mediator, allowed speed in a context in which it is difficult to manage time and resources,” continues the aid worker.
These devices allow communication with 90% of the world’s population and have been shown to have 96% accuracy
smaller than a smartphone and with a five-inch screen that fits perfectly in your pocket. These are devices that allow communication with 90% of the world’s population and have been shown to have 96% accuracy. “This is because they work with 10 intelligent context-specific translation engines. Which means that the more the speaker speaks, the more specific the interpretation will be,” explains Laura Méndez, spokesperson for Traductores VA, in a telephone interview. It took the company a year to develop this device, which can also translate handwritten documents. “The aid worker has to take a photo of the paper, the machine detects the language and then he only has to select the ‘Spanish’ option, if that is the language he wants to translate into,” says Méndez.
Between December and January, 57 migrants arrived at the Italian ports of Salerno or Civitavecchia in one of the missions and 115 in the other. In total, 172 people had artificial intelligence at their disposal to facilitate this trip. “This is progress because the more you take care of communication, the more you humanize the situations of the displaced, and ultimately, the more you take care of them when they get on the boat after being rescued,” Di Si concludes.
More smart technology
These translators are not the only smart technology initiative that is being applied to benefit humanitarian aid. In 2017, the UN Refugee Agency (UNHCR) developed the Jetson Project, an AI tool that was responsible for identifying areas of drought and violent conflict in Somalia by satellite, key factors in forced displacement in the African country. This method focused on predicting Somali arrivals to facilitate NGO preparation on the ground and thus allow for more preventive actions.
To facilitate the work of social workers and increase successful outcomes, the German Federal Office for Migration and Refugees – the BAMF – also developed “a pioneering system”according to him Migration Data Portal. It consists of increasing the efficiency of asylum procedures thanks to a software integrated identity management system that works with BAMF staff to register asylum seekers and collect information on their country of origin and travel history.
The new European AI law specifies that artificial intelligence is a “high risk” tool if used for migration, asylum and border control management
Another tool that has served as support in migration is the application GeoMatchwhich was released in 2020. It is an AI created between Switzerland and the United States that, through an algorithm, assigns refugees a destination selected from all the resettlement places available to them and compares them, offering the one with the best results of integration in each case. Migration Data Portalfunded by several governments and international institutions, alleges that GeoMatch has achieved an increase between 40% and 70%, in the employment rates of migrants, “if it is compared with current destination assignment practices, which are around 30% successful.” This happens because the system looks at specific information about the refugee, for example her language skills. The application thus builds a set of learning models that predict the expected work success of migrants based on their characteristics. The next step is called the mapping stage, in which the destination assignment is made by calculating the expected probability that at least one refugee, within a group, will find employment in the place in question.
A “high risk” tool
In January the new European AI law which specifies that artificial intelligence is a “high risk” tool if used for the management of migration, asylum and border control. For example, when the Administration carries out automated examinations of visa applications. Claudia Finotelli, an expert in control policies and specialized in visas, assumes that “these examinations are done based on data that already exists.” Over the past 20 years, rejection rates for three-month temporary visa petitions in European countries have been high. “The balance will continue to be high in visa applications that are studied by AI because the algorithm is replicated and the pattern is repeated,” Finotelli emphasizes in a call to this medium. According to her, “there is inequality at the regional level, with the African continent being the one that accumulates the highest rejection rates.”
A report carried out at the Complutense University of Madrid ensures that visa requests from countries like Nigeria are rejected in 66.4% of cases
A report carried out at the Complutense University of Madrid by Finotelli ensures that visa requests from people arriving from African countries such as Nigeria are rejected in 66.4% of cases; If they come from Gabon, they are denied by 61.7% and if they come from Guinea-Bissau, by 52.4%. This trend confirms other studies that relate how visa policy “has never been exclusively a tool for preventing irregular migration,” Finotelli points out, and that the general trend in the supply of short-term visas “highlights an asymmetrical visa regime, increasingly more open to Eastern European countries, although notably rigid throughout the Mediterranean.” Migrants from Asian countries, such as Pakistan (52.8% of applications denied) and Iran (31.9%), also registered a high rejection rate in 2022, the expert clarifies.
According to Finotelli, throughout the immigration cycle, from leaving the country to arrival and possible expulsion, “the Administration can use chatbots.” A chatbot is a computer program that uses artificial intelligence and natural language processing (NLP) to understand questions and automate responses, as if it were a human conversation. “It recognizes identities, performs behavioral analysis, offers psychological support… But the information is not always adequate,” she maintains. Given this, the European Commission, with the aforementioned AI law, is a pioneer in protecting the evolution of artificial intelligence to prevent racial biases from occurring and encouraging good practices of this tool.
You can follow Future Planet in x, Facebook, instagram and TikTok and subscribe here to our newsletter.
#doubleedged #tool #migrants