The personal assistant, endowed with artificial intelligence to handle sometimes tedious phone calls on your behalf, has set its sights on call centers. Well, not exactly it, but its creator has. Since this is not about a lonely carpenter Geppetto who, out of boredom, creates Pinocchio, but about Google whose AI puppet named Duplex potentially has access to all offices, homes, and smartphones in the world, the story becomes somewhat more serious.
A few months ago, Google introduced Duplex, an AI assistant that calls restaurants, hairdressers, and similar places on behalf of its users to arrange appointments or place orders. One of its qualities is convincing human speech, which has deceived everyone who has spoken with it. Google’s ‘puppet that lies’ reminded us of a modern and real version of a character from children’s stories who eventually becomes a real boy, just as Google Duplex is likely to become everyone’s reality.
>>> Google Duplex, the AI that stirred social ethicists with a haircut order
It started testing how Duplex can make it easier for users to make reservations, and according to fresh reports, Google is now looking for ways to integrate the system into customer service operations. The Information reports that this giant company is now focused on making its AI the employee of the month in call centers. Every month. Indeed, Duplex is expected to become the ultimate replacement for employees in call centers.
A source close to Google told the portal alphr that the company is in negotiations with a large insurance company where they want to implement Duplex to take customer calls. Google denied this, stating that its AI is not currently being tested in any company, but that they are currently focused on consumer use.
“We are not currently testing Duplex on companies as clients. Duplex is designed to operate in very specific situations, and we are currently focused on testing appointment reservations in restaurants and hair salons. The number of users testing the product is limited. It is very important for us to gain good experience, and we have chosen a slow, measured approach in which we take into account everything new that the AI has learned, as well as feedback from our testers.”
