Amazon Alexa Patent Will Record Even Before Wake Word
Los Angeles – If you’ve already had concerns about your privacy when it comes to Amazon’s Alexa devices or other smart speakers, don’t read this. Presently, Amazon’s smart speaker devices are engaged when you say “Alexa,” or whatever you set the wakeword to be. For example, if you said, “Alexa, shuffle Adele,” the speaker would start recording what you said after “Alexa” to get the information needed to process your request. This means that the device will only send your recorded voice to Amazon’s servers after you activated the device. Functioning under these restrictions provides a relatively formal system of communicating with a machine.
With efforts to improve Amazon’s customer experience, the company is apparently considering a change to the current system that would allow users to employ the device without the current restriction of needing to use a wakeword before each command.
The vision for this system was revealed when a patent was made public that lays out the plans for the new technology. The company has promised that this new method has not been enabled in any way so far. The release of this patent revealed a drawback to the current technology of smart speakers. If you started a command before saying “Alexa,” the speaker would not be able to register the information or process your request. Alternatively, Amazon is proposing a method that allows the word said before or after the command with identical results.
For instance, if you said, “Shuffle my workout playlist, Alexa” presently Alexa would not play the playlist because the command was said before the wakeword. With the technology laid out in the patent, however, when Alexa identifies the wakeword, the device will start to look backward to retrieve a command said before the word, using speech pauses as indicators. This system is possible because the smart speaker will be recording consistently, intending to delete what it deems unnecessary. Due to expected privacy complaints, the patent additionally addresses these concerns stating that all captured speech is not sent to Amazon’s servers.
A spokesperson for Amazon published a statement that reassured users that the technology in the patent is not in use and that potential use of the technology set out in the patent is purely speculative. The representative also added that many of the ideas in patent applications never turn into customer-facing products.
As Amazon continues to develop ideas to make devices more user-friendly and natural, this patent is likely an attempt to create a better product for consumers. It seems possible that at some point Alexa could develop an ability to know when it is being spoken to, so that a wake work is no longer even necessary. Although, if your primary concern is privacy, these changes are not likely something you look forward to.