While Apple’s HomePod has seen a gradual growth in sales in recent months, that growth is still being outpaced by the competition from Amazon’s Alexa and Google’s Home speaker. This is almost certainly due in no small part to the lower price tags of those competing speakers, but it also doesn’t help that the HomePod has generally lagged behind in its voice assistant capabilities due to Apple’s choice to emphasize the speaker quality rather than Siri’s role in the device.
We’ve previously discussed some of the things that we think Apple needs to address in the HomePod to make it more than just a speaker that also happens to support Siri, and it looks like Apple may actually be addressing some of these things when iOS 13 arrives later this year.
Tucked away in yesterday’s Bloomberg article about the big app changes coming to iOS 13 was a note that Apple is also “planning to let the HomePod speaker respond to different users’ voices, creating a much requested multi-user mode.” This would address some of the biggest complaints about the usefulness of HomePod in a family setting; the voice assistant not only can’t distinguish who is talking to it, but Siri is also far too easily triggered even when nobody is talking to her.
Preventing False Activation
In fact, it’s actually quite surprising that Apple hasn’t at least offered some “Hey Siri” training to its smart speaker. The capability of training an iPhone to only respond to the owner’s voice has been around since iOS 9 was released in 2015 to support the “always-listening” implementation of “Hey Siri” made possible by the new iPhone 6s’ M9 co-processor — a necessary feature to reduce false triggers once Siri could be called for even when an iPhone wasn’t plugged into a power source. The feature has been a staple on all of Apple’s iOS devices since, and was most recently even added to its second-generation AirPods.
However, despite Apple acknowledging the need for this — and highlighting how the Hey Siri feature works on more than one occasion — Apple’s smart speaker has remained relatively “dumb” when it comes to knowing who is talking to it, or even when it’s actually being called for; it’s not uncommon for Siri to be randomly triggered by whatever is playing on a TV in the same room, even when it sounds nothing like “Hey Siri.”
It seems that Apple chose to avoid adding voice training to the HomePod simply due to the number of people in a household that could be talking to it, perhaps to avoid the hassle of training every family member, or to accommodate visitors who might also want to make Siri requests. While it’s less clear how iOS 13 may address this — voice recognition by itself doesn’t mean that HomePod may not still support “anonymous” voices, or some kind of “guest mode” but we’d certainly like to see improvements to help ensure that Siri only comes up when she’s called for.
Multiple User Support
The real advantage to voice recognition, however, will be in multiple user support for the HomePod. While the smart speaker assistant has supported personal requests from the very beginning — the ability to send and read messages, set reminders, create and add to notes, and make phone calls, and later manage calendar events — this only works for a single user. This not only leaves other family members out of the fun, but also makes pronouns confusing: “My” will always refer to the primary user, so if another family member says something like, “What’s on my list for today” they’re going to get the primary user’s reminders read back to them. This is also exacerbated by the fact that the HomePod is more likely to respond to “Hey Siri” than the users’ own iPhone or iPad.
There’s not much known about how Apple will go about implementing multi-user support. We’d like to see the HomePod gain the ability to query iCloud directly, since the data is already stored there anyway, although we suspect that Apple will simply expand the current system to support multiple users’ iOS devices, requiring that user’s paired iPhone or iPad to be on the same Wi-Fi network in order to answer their requests.
This may not only be simpler, but also offers a bit of security as we’re not convinced that Apple is ready to rely on voice recognition as an authentication method. Requiring that the user’s primary iPhone (or iPad) be on the same Wi-Fi network will offer some degree of privacy. Further, a HomePod user can already choose to require authentication directly on their iPhone when requests are made for more personal information, such as asking Siri to read notes, so this could be handled in the same way for multiple users.
There’s also been no word on whether Apple will expand personal requests to additional categories; while SiriKit is supported for third-party apps and shortcuts, there are still things that the speaker version of Siri lacks the ability to do, such as checking email messages or reading audiobooks. We’d also really like to see HomeKit become more multi-user aware — a request like “Turn off the lights in my bedroom” can be very confusing in a family household.
Both Amazon’s Alexa and Google’s Home speakers have supported multiple users for over a year, with Google having the feature almost right out of the gate back in 2017, albeit with some quirky limitations. While it’s perhaps not surprising that Apple is behind the curve on this one, since the company prefers to do things “right” rather than “first”, we’re definitely excited to see how iOS 13 will improve the experience of using Apple’s smart speaker in family settings.