There has been a growing furore around the privacy of recorded voice commands used by digital assistants on various platforms when it has been revealed that companies routinely employ humans to listen to a percentage of these voice commands to improve the technology.
The latest revelation has been that this also extends to Xbox One voice commands, according to an article by Motherboard. Motherboard spoke to several contractors who report that they have been involved in analysing such recordings; which in most cases were intentional commands recorded by the Kinect camera, for example, but on other occasions involve accidental activations, and therefore included information or speech the users presumably thought was private.
This news is particularly significant because as a gaming console, the Xbox One is more likely than most voice-recognition technology to be used by children, for whom more strict privacy guidelines apply.
Microsoft for its part has been open about its use of humans to improve its voice technology, saying “we’ve long been clear that we collect voice data to improve voice-enabled services, and that this data is sometimes reviewed by vendors.”
It should by now be clear to any user who has been keeping an eye on the news, that wherever a voice-recognition technology is used, there is a very high likelihood that a person could be listening in and that this is a trade-off in part in creating a workable and usable voice-recognition service.
While Microsoft has been unapologetic about this practice, it should now behove users to take responsibility for their own privacy and enable or disable the technology or the feature as required. Also, knowing that people may be involved, it may be worth toning down the insults.