Siri was probably the last smartphone feature that convinced me to buy a new phone, even though the one I was using was working fine. In late 2011, when Siri was introduced, I had just accepted a job in Los Angeles, and was psyching myself up at the prospect of spending a lot of time in my car. The idea of having this little, helpful friend in my car that I could talk to — get directions, send texts, check traffic — while I kept my eyes on the road was tremendously appealing. Early commercials were enticing enough that around the same time that I bought a used Honda Civic, I grabbed an iPhone 4s.
In reality, of course, Siri didn’t really work like that. Text messages that went beyond a few words were usually garbled. Apple Maps was a mess and didn’t have the actual real-time traffic information you need to navigate Los Angeles, so I ended up just typing wherever I was going into Waze, regardless. Its knowledge base was thin; its understanding shallow. Connectivity was important: Sometimes, it seemed like Siri couldn’t connect to Apple’s main servers, rendering Siri apologetic and useless. I soon grew tired of Siri screwing up, and stopped using it, except occasionally to place a phone call while driving.
Things in the past five years should have gotten better, and in a lot of ways they have. Voice assistants on phones began to take off about the same time neural networks started being used, and voice recognition on Siri and its rivals — Google Assistant chief among them — is markedly better than it was just a few years ago.
Bixby, a new voice-activated phone assistant rolled out by Samsung with its Galaxy S8, will also soon be able to accept voice commands (voice recognition is currently unavailable, and it’s not clear when it will be). In a demo shown off to a small group of reporters, a Samsung employee rattled off somewhat complex operations in natural language that Bixby will ostensibly be able to follow, e.g., “Send all my pictures from Barcelona to my mother.”
Of course, the demos for these things are always impressive — people are always working out complex dinner plans on the fly, or filtering all their photos from a cool vacation. In the real world, Siri and Google Assistant work very well for certain things — setting timers, sending short text messages, asking simple questions — and are terrible at others.
Plus, there’s the social aspect of all of this. Perhaps in ten years, we’ll have ascended to Spike Jonze’s Her-esque levels of acceptance, where guys with very high-waisted wool pants can wander around talking to their phones in peace. But right now, using Siri or Google Assistant in public, and making sure it understands me, requires that I use a certain tone — semi-stilted and not quite natural — that just makes me feel like a weirdo. And I don’t think I’m the only one.
This isn’t a new problem; tech reviewers, and consumers, have been noting how awkward it is to use Siri and its peers for the last five years. What’s odd now is that there is a popular gadget that no one seems to mind talking to. I have an Amazon Echo that gets daily use in our house that I am quite fond of. It’s guilty of all the things I was just kvetching about. It doesn’t always understand what I’m trying to say; I have to talk to it in a weird voice, almost like I’m talking to a dog or a small child; and in many ways, it’s severely limited in what it can do compared to Siri or my Google Assistant. But I talk to the Echo a lot more, and it’s not just because I don’t feel like a weirdo. It’s because it’s actually in a place to be useful. We need something that can set timers for cooking; play music or the radio while we’re making breakfast or dinner; and order items off Amazon while I’m thinking about the fact that we don’t have any more Mrs. Meyers dish soap, instead of just hoping I remember later that day.
I’m not alone in this. Asking around, I found even generally tech-friendly people were in the same boat: They didn’t really use Siri or Google Assistant, except maybe to set a timer or alarm, and only in the privacy of their own homes. The one person I could find who said they used Siri a lot, a content strategist in Chicago, said it was mainly at home. When I asked him if he used Siri in public, he IMed back: “Sometimes, but it certainly feels weird. Especially if it fails to get the request right the first time. There’s something weird about repeating yourself to a voice assistant versus a human who can’t hear you over noise.”
So, then, a working theory on why I (and a lot of other people) don’t really want to talk to Siri or any other AI on our phones, but plenty of people have happily adapted to Alexa or a Google Home. There’s a social cost in the form of feeling like a dorkus malorkus when you use your phone’s voice assistant in public, and right now, those assistants can’t do enough useful things reliably and consistently to make it worth paying that cost. And because our phones offer another way to do those useful things without speaking — check train times, look up a recipe, caption and share a photo on Instagram — there may never be any inflection point in the future where we start talking to our phones as a primary or even secondary way to do most tasks. Alexa and Google Home may fall short on competency at times, but they both open up another way to do useful things that wasn’t really available before and don’t incur any social cost.
The smartphone suddenly gave you the digital world in your pocket, available at the tap of a finger. The AI home speaker suddenly opened the possibility of talking to your home in a way that felt both futuristic and familiar. But maybe we’re content with just sticking to tapping away on our phones in public, and only speaking to our gadgets in private.