At Google I/O 2018, the company unveiled a host of new features for its products and services. But if you’d ask us, then we’ll say that Google Duplex was the most interesting feature that stood apart. What it does is that it lets the Google Assistant make a call on your behalf for small things like making a salon appointment or reserving a table at a restaurant. Since then, a year has passed and now it is being reported that one out of four duplex calls is handled by a human.
According to a report by the New York Times, 25 percent of the times it was actually a human who answered the call instead of an AI bot. Now, according to Google, the company has revealed to the New York Times that a human answers 25 percent of the calls that are placed through Duplex and about 15 percent of the times, it is actually an AI robot that answers. The call is then handed over to a human at a later stage to intervene if that is required.
The report also mentions that Google Duplex was tested by calling several restaurants and in cases where a human handled the calls, the experience was not so satisfactory. But in cases where the Google Assistant did get it right and the call was handled by an AI bot, then it was very impressive. In one instance during a call with the manager at a Korean restaurant, the Assistant handled the situation smartly and answered to repeated queries with patience.
But despite Duplex working perfectly as the way it should (when an AI bot does answer a call instead of a human), the fact cannot be ignored that the service is still very much reliant on humans. Now, Google is not saying that this isn’t an issue but the company in its defence does add that, Duplex calls are still dependent on humans because the company wants to take a conservative approach to be respectful towards a business. Additionally, the company also states that human involvement will be required in calls where the Assistant is not able to decide whether the restaurant is taking a reservation or not. Gradually Google will work on improving the service and reduce the reliance on humans in the coming future.