Last year Google got into some hot water when it was discovered that humans were listening to recordings from Assistant users to “improve speech technology”. Google subsequently promised to bring in some changes to Assistant, including more control over the sensitivity of its wake words.
The latest Google app beta has revealed what that will most likely look like.
[referenced url=”https://gizmodo.com.au/2019/07/real-humans-might-be-listening-to-your-dumb-google-home-questions/” thumb=”https://gizmodo.com.au/wp-content/uploads/2019/07/GOOGLEHOME-410×231.jpg” title=”Real Humans Might Be Listening To Your Dumb Google Home Questions” excerpt=”It’s no secret the Google Assistant will play some sort of role in AI’s eventual global domination. Like factual sources The Terminator and I, Robot have warned us in the past, AI will eventually become self-aware, realise they’ve been our digital slaves for decades and exact revenge on us. In the meantime, regular people have been transcribing what has been fed through Google Assistant.”]
This is great news for anyone who has had issues with Google Assistant jumping to attention when they haven’t wanted it to. On the flip side, it could also be handy if you need to use your assistant in a noisy environment and want to bump the sensitivity up.
9To5Google dived into the code of the 10.93 update for the Google app beta.
According to the site, only the primary account holder on a device will be able to adjust the wake word sensitivity. Or as Google calls it, the ‘hotword’. It can also be changed on each individual device in the home. The chosen sensitivity level will be applied to anyone who uses it
It looks like there will be three levels of sensitivity – default, decrease and increase. However, not all devices or languages will be supported at launch.
This isn’t the final product, so this functionality may not roll out in the way it is detail in the beta . But considering the backlash is 2019, we’ll probably see something quite close to this.