User privacy has been a big deal in recent years and with companies like Google and Apple tracking your every move, it’s hard to get the sense of secrecy and isolation that people used to get in the past. Unfortunately, it’s going to take a turn for the worst soon.
According to a recent patent filing by Amazon (via Gizmodo), the company is planning for something it calls “system for capturing and processing portions of a spoken utterance command that may occur before a wakeword”. This process basically means that the user won’t have to say the wakeword, which by default is Alexa on the Echo devices. Currently, Echo devices will only work if the user says Alexa, followed by the command like “Play the music” or “Turn on the lights”. However, with the new technology Echo will wake up after the wakeword but will be able to rewind back and detect the start of the command to execute it. For instance, if the user says “Play the music, Alexa”, it would be able to rewind back to the start of the sentence and execute the command.
While this technology would certainly be useful and will improve the smart speakers, it would also mean that Echo would be recording and listening to the conversations all the time. The patent does address the privacy concerns by deleted all the unused recordings and not sending any of it to the Amazon servers. Amazon has confirmed that the technology is currently not being used by the company but at the same time didn’t confirm if Amazon will ever use it. This new patent would certainly improve Alexa and would make it more natural but will be a big invasion of privacy in a way. That said, privacy is a burning topic right now so it might be too soon to tell if the technology will be adopted by the customers or not.