Imagine an Amazon Echo that can respond to any commands at any time, even if no wake word was used. Instead of saying, "Hey Alexa, Play this song," a person can drop the Alexa altogether and bark a command as they would to, say, an actual human being.
Apparently, Amazon is imagining the same thing. A just-surfaced patent now reveals that the company is considering erasing "Alexa" from the equation to make the process of asking commands more natural, smooth, and streamlined.
No Need For Wake Words?
In essence, such a feature would allow Echo devices to start picking up audio even before the user utters their wake word. The patent states that it would allow users to start more natural conversations with their devices. At the moment, Echo devices are simply unable to pick up commands until the user says their chosen wake word.
So, how would this feature work? In practice, Alexa would be "looking backward" at recent things said aloud prior to hearing its name. In a command along the lines of, "Play this song, Alexa," the device would "wake up" at the mention of "Alexa" then go back a bit to determine the prior phrase said. Obviously, to accomplish something like this, Alexa would have to be listening all the time, storing and processing speech, then quickly erasing recordings if deemed not relevant.
Privacy Concerns
It's not difficult to imagine how such a feature, if implemented, might make Echo owners even more concerned about privacy than they already are now. There's already a considerable amount of pushback with regard to how Amazon records, stores, and uses Alexa recordings to collect user data. In many occasions, Amazon has already demonstrated that Alexa recordings aren't as private as users think — it was reported recently that a team of humans listen to and process Alexa commands, and those people might have access to potentially identifiable information, such as location data.
A feature like this stands to make such woes even worse. Clearly, Amazon thought of possible repercussions, which is why the patent states users would be given the choice to allow Alexa to record and store audio for between 10 and 30 seconds at a time.
Amazon has made clear that the technology described in the patent is not currently in use.
"Like many companies, we file a number of forward-looking patent applications that explore new scientific ideas that may not make it into customer-facing products," a spokesperson for the company told Engadget. "Patents take multiple years to receive and do not necessarily reflect current or near-future state of products and services."
