Voice Recognition Technology To Detect Origin Of Refugees In Germany

Author Photo
Mar 17, 2017
10Shares
Submit

According to a report by Die Welt, Germany will soon employ voice recognition technology to find out the origin of refugees. The technology recognizes the dialect of the refugees and detects the region they are most likely to belong. Authorities from the Federal Office for Migration and Refugees (BAMF) also believes that the tech will help dialect experts in easing out the numbers and make the process speedier.

Last year, Germany witnessed a drop in the number of refugees to 280,000 from 890,000 in 2015. But, the backlog of the asylum applicants is still 430,000 and that is why the authorities are seeking external help in the form of voice recognition policy. Banking and Insurance industries have been using voice recignition based softwares to identify refugees. The software analyzes the dialect of refugees on the basis of registered speech samples. It will help the authorities to find out of the refugee is coming from a war affected region and truly needs shelter.

germany-surveillanceRelatedSurveillance States: Germany Wants Tech Companies to Be Legally Required to Offer Backdoor Access

Germany will initiate the software tests in two weeks from now, and if it is found to be accurate then the authorities will start using it on a wider level by the year 2018. In Germany, half of the refugee population has got the asylum while around sixty percent refugees have got other types of protection services to keep them away from facing deporatation. Among these refugees, most come from war-torn regions like Syria, Afghanistan, and Iraq.

The use of the software sounds great for lessening the backlog burden on the authorities, but the experts are concerned about the accuracy and integration of the tech in the process. Also, it would be a bit tough to train the immigrants to react to the tech and record their voices clearly. A large part of speech recognition also depends on the clarity of the tone, so yeah, there will be instances when the results would fail to match.

University of Essex linguistics Professor Monika Schmid told Deutsch Welle :

I don’t see how automated software can distinguish whether a person uses a certain word or pronounces it in a particular way because this is part of their own repertoire or because they were primed to do so by the interviewer or interpreter, she said. Identifying the region of origin for anyone based on their speech is an extremely complex task. Both humans and machines can easily be wrong, but humans are probably better at realizing this.

Putting her trust in the manual work, Schmid believes that Authorities from the Federal Office for Migration and Refugees (BAMF) should leave the job on manual detection of the dialect by 45 linguistics experts who are trained in 80 languages and have an experience of dealing with such processes.

Submit