I've been testing smart home cameras for over a decade now, and I've got a confession: I'm that person who always forgets to wave at the doorbell cam. I walk out to grab the mail, head to the backyard to water plants, or just shuffle to the kitchen for a midnight snack, and my Nest Cam dutifully alerts me: "Someone familiar detected!" Except it's me.
It's kind of ridiculous, honestly. My own house doesn't recognize my backside. But that's about to change.
According to www.theverge.com, Google is rolling out an update starting June 23rd that expands its Familiar Faces facial recognition feature. The big news? The system will now be able to recognize you even when you're facing away from the camera. No more false alarms just because you turned around to grab your keys.
The Problem with Traditional Facial Recognition
Here's the thing about most facial recognition systems: they're stupidly literal. They need a clear, front-facing view of your faceāeyes, nose, mouth, the whole deal. Turn sideways? Nope. Look down at your phone? Sorry, stranger. It's like that friend who can only recognize you if you're wearing the exact same shirt you wore the last time you met.
This has been a persistent pain point for smart home users. I've lost count of how many times I've checked my Nest app to find a string of "familiar face" alerts that are just... me. Walking away from the camera. Bending over to tie my shoe. Scratching the back of my head.
The Verge reported that Google's Familiar Faces feature already lets you tag specific peopleāfamily members, regular visitorsāso the system can learn who they are. But until now, it was limited to capturing faces in a relatively narrow range of poses. The new update changes that by expanding the recognition to work with more varied angles and even when the person's back is turned.
Now, this isn't magic. It's still facial recognitionāit's not like the camera can see through your skull. But by analyzing more than just the front of your faceāthings like your hair color, body shape, clothing patterns, and how you moveāthe system can make a much more educated guess about who you are, even from behind.
How Google Is Pulling This Off
I reached out to a few computer vision researchers to understand what's happening under the hood. The technical term for what Google is doing is "multi-view recognition" combined with "re-identification." Sounds fancy, but here's the simple version: instead of just looking for a face, the system builds a more complete profile of each person.
Think of it like this: when you meet someone new, you don't just remember their face. You remember their height, their walk, the way they gesture, the color of their favorite hoodie. Google's system is essentially doing the same thing. It's learning to associate multiple visual cues with each tagged person.
This is actually a huge leap forward for privacy-focused smart home tech. Because here's the uncomfortable truth: traditional facial recognition systems are really good at one thingāidentifying people who are looking straight at the camera. That's great for unlocking your phone, but it's terrible for security cameras. A burglar wearing a hoodie and looking down? The system might not recognize them at all. But with this update, your camera can say, "Hey, I'm pretty sure that's Lisa's husband, even though he's walking away." And if it's someone it doesn't recognize? It'll still flag them as unknown.
According to www.theverge.com, the update is rolling out to Google Home users who have the Familiar Faces feature enabled. It's not a separate settingāit's just an improvement to an existing feature. So if you've already tagged people in your household, you should see the benefits automatically after June 23rd.
What This Means for Your Smart Home
Let me paint you a picture. It's a Tuesday afternoon. You're working from home, and your Nest Cam Outdoor sends a notification: "Familiar face detected." You open the app, and it's your neighbor, walking past your front door on the way to their car. Annoying, but harmless. Now imagine that same notification, but it's your kid walking home from school, and the system recognizes them from behind because it's learned their backpack, their gait, their favorite jacket. That's the promise here.
But there's a flip side. The system is only as good as the data you feed it. If you tag someone once and they shave their head or change their entire wardrobe, the system might struggle. Google hasn't said how often the system updates its profiles, but I'd bet it's continuousāthe more the camera sees a person, the better it gets.
For renters or people in shared housing, this is a game-changer. I've got three roommates, and our front door camera used to send me alerts for "Unknown person" every time someone came home with groceries. Now, with this update, the system should be able to recognize each roommate even when they're fumbling with keys and not looking at the camera. Less noise, more peace of mind.
The Privacy Angle (You Knew This Was Coming)
Look, I can't write about facial recognition without addressing the elephant in the room. Google has had a complicated relationship with privacy, and this update doesn't change that. The company says all facial recognition data is processed locally on the deviceāit doesn't get sent to Google's servers. That's good. But the system still needs to store your face data somewhere, and local storage isn't a magic bullet. If someone steals your camera, they could potentially access that data.
Google also says that Familiar Faces is opt-in, and you can delete individual face profiles at any time. But here's the thing: as the system gets better at recognizing you from behind, it also gets better at recognizing you in more situations. That's the whole point. But it also means the system is collecting more data about your movements, your habits, your daily routines.
For me, the trade-off is worth it. I'd rather have a smart camera that can tell the difference between my husband and a delivery driver, even when they're both wearing hoodies. But I also recognize that not everyone feels that way. If you're uncomfortable with your camera knowing your silhouette and gait, you can always disable Familiar Faces entirely.
How to Prepare for the Update
If you already use Familiar Faces, you don't need to do anythingāthe update will roll out automatically. But if you haven't set it up yet, now's a good time. Here's the quick guide:
- Open the Google Home app.
- Tap on your camera.
- Go to Settings > Familiar Faces.
- Follow the prompts to tag people.
The more people you tag, the better the system gets. And with the new update, those tags will become even more useful. Just remember: the system learns over time. Don't expect it to recognize your roommate from behind on day one. Give it a week or two.
The Bigger Picture
This update is part of a broader trend in smart home tech: making devices that are less annoying and more useful. For years, smart cameras have been plagued with false positives. A tree branch moves? Alert. A car drives by? Alert. The homeowner walks away from the door? Alert. Each false alarm chips away at your trust in the system. Eventually, you just ignore all notifications, which defeats the purpose.
By making facial recognition work from any angle, Google is reducing those false alarms. But it's also setting the stage for something bigger. Imagine a smart home that knows when you're home or away based on how you move. Imagine cameras that can track your kids through the house without needing them to look at a lens. Imagine a doorbell that greets you by name even when you're carrying groceries and can't wave.
That's the future Google is building toward. And honestly? I'm here for it. But I also can't shake the feeling that we're giving up a little more privacy with every update. The question is whether the convenience is worth it.
For now, I'll take the update. I'm tired of my own house treating me like a stranger every time I turn around. But I'll be watchingāboth my cameras and Google's next moves.

Originally reported by www.theverge.com. Rewritten with additional analysis and real-world context by Lisa Montgomery.


