Has anyone else been told off by the Assistant when it overhears swearing? And is there anything I can do to stop it, short of censoring myself while I’m speaking in private?
Today, I asked the assistant to turn on a light. Once the light was on, I saw that I had stepped in food and tracked it across the floor, and I let out a “What the ****?!” (The deleted word is a four-letter, garden-variety swear that starts with the letter F.)
Evidently, the Assistant was still listening. It responded by addressing me by name, and telling me that although it is only a virtual assistant, it’s not acceptable for me to use that kind of language. Then it said something about how there are better ways to get product support, and asked me if I needed to send feedback to Google about the performance of the Assistant. Well, I didn’t before, but now I do. I used a harmless profanity in an otherwise empty room of my house, directing my comment at no one — not even the Google Assistant — and a smart home speaker told me off in terms that suggested that I had done something terribly inappropriate.
This feels really intrusive, and I hate it. I know there is a way to change whether the Assistant will speak profanities that it encounters in titles and search results. Is there a way to stop the Assistant from telling me off when it picks up incidental swearing? I am generally happy with the performance of this thing for controlling smart home devices and playing music, but I don’t appreciate the pearl clutching when the Assistant hears an F-word that wasn’t even directed at it.
The Google assistant is designed to encourage respectful communication amongst different communities and families around the globe. If it changed this basic manifesto I wouldn't use it anymore myself.
Please leave your feedback via one of the options within the link below as this is the best way to communicate with the engineering team.
Thank you for responding and for the feedback link. I will make sure I use it.
I think I understand what you mean about the way the product is designed, but I haven’t been disrespectful of anyone. In fact, at the time that this occurred, I was in an empty room and I wasn’t even addressing the virtual assistant. I don’t think it’s right for Google, or for anyone for that matter, to decide whether or not it’s okay for me to privately use a swear word while I’m alone in a room inside my own home. (Also, bear in mind that we are talking about the f-word, here. We are not talking about a slur.)
Would I continue using the speaker if it stopped doing what it did today? Emphatically, yes. For me the issue is that the assistant didn’t need to respond to what it heard. The utterance that it responded to was not a question, a direction, a request, or an insult. I also was not attempting to speak to any other people — I wasn’t making a broadcast, making a call, or sending feedback. The response it gave doesn’t make sense in context, and it added nothing, but it also managed to feel very intrusive and judgmental. It’s a net negative.
For whatever it’s worth, the feedback I intend to give after following the link includes two points:
1. The assistant needs to do a better job of determining when someone is speaking to it, and when it has picked up speech incidentally. It needs to be better about not making a lengthy, disruptive response to speech that was not directed at the assistant. Based on the actual words that I said when it took the time to admonish me today, it should have been obvious to the assistant that my exclamation was not directed at it.
2. When the assistant does think that someone is swearing at it, the lengthy admonishment it provides in response can be too much. Obviously I am not advocating for using the speaker in order to yell profanities at someone, but it’s not the same if someone is talking to the speaker itself. The speaker is a thing. The virtual assistant is not a person. Some people like to swear! (And some find themselves letting out a swear word every now and again completely by accident, such as when they have trod in a smooshed Nutrigrain bar.) Even if I had been deliberately swearing at the speaker, there isn’t anything inherently wrong with me swearing at an object in my house, particularly when I am not using the device to speak to other people.
I hope that this is something that gets updated and improved.
No offense, but why are you defending yourself or framing your complaint in parameters, like if you had used a racial slur in the privacy of your own home, then it would be appropriate for Google Assistant to tell you off? It's bad programing on Google's part, plain and simple.
I agree that this is an instance of bad programming. And in general, I don’t think anyone has any business telling me what to say in the privacy of my own home.
The caveat, however, is that it is possible to interact with the Google Assistant in a way that communicates with real people — for example, when sending feedback to engineers. I think that it would be reasonable for Google to prevent its products from being used in a way that is abusive of its own employees, so the context here does still matter. Bottom line, either the assistant misread the context, or it has been programmed to have a particularly strong response to uses of the f-word regardless of context.
There is a setting in google home assistant "Continued Conversation". If this is on, it allows for follow-up questions/instruction without repeating "hey Google". That's probably what's caused this.
What you wrote is exactly how I feel. Thank you, TrekTrek. I know it's just a machine, but I feel really incensed at having my behavior censored by a robot. I have no moral obligation to a machine. None. I'm actually surprised at how much this bothers me. I just experienced it today. I said f... something and got exactly what you wrote above.
Lol it would explode from over use in my house if it had to correct us every time.
But on a serious note. This way of speaking is part of normal conversation in ireland so it can **bleep** off if it thinks it's acceptable to tell me how to speak in my own home!
We appreciate people who are keen on sharing their thoughts with us. We're always looking for ways to improve and we'll take this as feedback. Let me know if you have other questions or concerns, otherwise I’ll be locking this thread after 24 hours.