Company develops AI voice filter that turns angry screams into calm speech

A Japanese company has developed an artificial intelligence filter that can detect angry shouts and translate them into calm speech to reduce the stress of call center employees.

Big tech companies have long been aware of the stress their call center employees face every day, as many customers take out their anger and frustration on them. Some of them have even introduced stress management programs that include relaxation techniques, meditation, yoga, or therapies for irritability and anxiety. But one Japanese company may have found a much more efficient solution – by using artificial intelligence to completely eliminate shouting and aggressive speech from the system. SofBank claims to have spent three years developing a speech filter that detects shouting and automatically translates it into calm speech.

Photo: Unsplash

“We developed the emotion suppression system in response to the societal problem of customer harassment of call center employees and to protect them,” said Toshiyuki Nakatani of Soft Bank, one of the developers of the innovative AI filter.

SoftBank’s speech filter consists of two phases: in one phase, the AI ​​detects an angry voice and extracts key points of the speech, and in a second phase, it uses acoustic tools to transform it into a more natural, even polite tone. Interestingly, the filter does not change any of the words spoken by the person, but rather significantly softens the intonation. The call center agent still hears any insults expressed, just in a gentle tone, which should help reduce stress and anxiety.

To train the AI, SoftBank engineers asked ten actors to record at least 100 common phrases, including shouting, accusations, threats, and demands for an apology. In total, more than 10,000 voice data were used to train the AI ​​filter.

It’s unclear when SoftBank plans to implement the new scream-filtering AI in its call centers, but it will be interesting to see how well it works and what impact it has on the emotional well-being of operators.