In the future, Siri would be able to unlock your device while obeying one of your commands by simply recognizing your voice. While that may seem an impossibility at the moment, it may not be for much longer as one of the many patents that were given to Apple and revealed after their publishing in the U.S. Patent & Trademark Office on Tuesday included “Device access using voice authentication,” according to reports from iDrop News.
Potentially, Voice ID would be the system that recognizes a user’s voice command. According to the report, the Apple device would recognize the voice of its registered owner through “a ‘voice print’ comprised of multiple examples of the user’s speech.”
In a hypothetical situation, a user could give its device a request such as “Siri unlock my iPhone” and then through Voice ID have their phone unlocked remotely. According to the report, while the name Voice ID is not official, that is essentially the technology that has been patented.
It could even go beyond a request to unlock the device, with the device being unlocked due to the vocal recognition, ending the Siri response of “You’ll need to unlock your iPhone first.”
While a vocal recognition system at first glance would appear to be quite simple to crack, it appears that Apple has already developed a system to remove any chance for a false positive.
— iDrop News (@iDropNews) October 16, 2018
According to the report, Apple will have developed a “voice-print” for each user through “statistical models of the characteristics of a specific person’s pronunciation including voicing, silences, stop bursts, nasal or liquid interference, frication, and more.”
The patent also reveals that the entire system will have no text aspect at all, removing any chance of a spoken password or command to unlock the device.
If the technology is achieved, when a device hears a voice command, it would compare said command to its voice-print. If the voice it receives matches the print, the device would automatically unlock and complete the command.
The report also suggests that the patent could account for a second layer of authentication before any user’s sensitive information is revealed through voice recognition. This would allow users that are not the device’s registered owner to still be able to use the device to answer general questions, with the technology able to understand that the user is not the owner and thus withhold any valuable information when responding.
In situations where the voice can’t be properly authenticated, the device will likely request some form of password or another form of validation, similar to how Touch ID or Face ID responds to failures.