An artificial intelligence created to “westernize” the accent of call center employees

In Silicon Valley, the young company Sanas has developed an artificial intelligence aimed at “Westernizing” accents, reports this Thursday, August 25, the British newspaper The Guardian.

The device is especially aimed at call centers in Africa or Asia, where employees are regular victims of disrespect from customers, according to Marty Massih Sarim, current head of the company.

At the moment, the artificial intelligence is fixed on the American voices and therefore only in English. Its purpose is to soften and hide accents. The effect is immediate and “undetectable” to the person on the line.

Can be used as desired.

Sanas says his system is already used by more than 1,000 call center workers in the Philippines and India. He specifies that employees are free to use it or not, depending on their choice.

A large number of Western companies have moved their call centers to countries in Asia or Africa. “Obviously, it’s less expensive to handle these calls in countries other than the United States. […] That’s why all the work is now outsourced,” says Marty Massih Sarim, president of Sanas and a former call center employee. As a result, employees are sometimes faced with irate customers.

In addition, Sanas indicates that many companies ask their employees to practice speaking “American.” Therefore, artificial intelligence aims, in the eyes of its creator, to avoid this additional workload.

Hide accents, but not discrimination

But the practice worries. Aneesh Aneesh, sociologist interviewed by The GuardianHe talks about his fears. In the short term, avoiding language training for employees seems like a good thing. But in the long term “it is a danger”, she believes. The sociologist fears that this will lead to “indifference to difference.”

“Like many systems presented as solutions, this one does not take dignity or humanity into account,” laments Chris Gilliard, a researcher specializing in the effect of technology on marginalized communities, for the British newspaper. “It will only support people’s possible racist and discriminatory beliefs.” “Loss of individuality”, “forgetting the other”: the researcher is formal, for him, this solution is not one.