At a conference on artificial intelligence, the scientific leader of Amazon’s voice assistant, Rohit Prasad, said that a body would allow the voice assistant to gain intelligence.
A body to better understand the world
As MIT Technology Review recalls, “Alexa has many advantages over a human brain – like access to a vast encyclopedia of useful facts.” This allows him to determine if the user is talking about a person, a place or an object. On the other hand, voice assistants reach limits by the absence of physical links with the world on which they must provide information. Their answers are sometimes unsuited to the situation because they lack discernment. “The language is complicated and ambiguous by definition. Reasoning and context are important, “says Rohit Prasad. Eyes could thus allow Alexa to understand non-verbal language, which is very important for communication. The idea is to teach him to do certain tasks that are obvious to humans, but that requires understanding a complex context for a machine.
Amazon would like to go further
After integrating Alexa into more than 100 million connected objects, Amazon would like to go further. The American giant could indeed give a body to his intelligent assistant and perhaps make a real robot. “The only way to make PDAs really smart is to give them eyes to explore the world,” said Rohit Prasad, vice president of Amazon, who leads the team of scientists who in Alexa point, at a conference organized by the MIT Technology Review (MIT Technology Review) on March 26. He considers that the assistant of Amazon would be more efficient with a body with cameras to move and observe the world around him.