Thank you for subscribing!

'These types of emergencies can be reversible — but only if someone knows the emergency is happening'

Google has unveiled its all-new Pixel Watch 5, which arrives with a one-of-a-kind health feature that could save lives.

Called breathing emergency detection, it's Google's newest Artificial Intelligence (AI) tool that can tell if you're having trouble getting enough oxygen. For instance, if you're experiencing choking, an accidental drug overdose, or a severe case of pneumonia that causes your blood oxygen levels to plummet dangerously low, the watch picks up on it.

And if you become unresponsive and can't call for help yourself, it steps in and automatically contacts emergency services on your behalf.

Pramod Rudrapatna, Senior Product Manager at Google Health, said in a blog post: "These types of emergencies can be reversible — but only if someone knows the emergency is happening, intervenes, and calls for help. But many of these emergencies happen when there’s no one around."

The watch uses an AI-powered algorithm that pulls together data from its heart rate sensor, motion detector, and altitude sensor to spot when something's gone wrong with your breathing.

To make sure it doesn't accidentally trigger when you're just sitting still on the sofa, it checks that you're in a low-movement state without any sudden changes in altitude.

The system also makes sure you haven't been asleep for more than 30 minutes.

It's specifically designed to filter out false alerts from conditions like sleep apnea, where oxygen levels can dip during sleep without it being a full-blown emergency.

However, if the watch thinks you've stopped breathing, it doesn't just jump straight to dialling 999. There's a step-by-step process designed to give you every chance to cancel if it's a false alarm.

First, you'll feel a strong vibration on your wrist alongside a full-screen alert on the watch face. If you don't dismiss it, an audible alarm kicks in and a 30-second countdown starts.

If there's still no response, the watch then uses your connected phone — or its own LTE connection — to ring emergency services automatically.

Once the call connects, an automated voice message tells the operator that a breathing emergency has been detected, that you're not responding, and shares your exact location.

If you've set up emergency contacts, they'll also get a message letting them know you might need help.

To be able to pack this life-saving new tool into a wearable, Google brought in internal medicine doctors, cardiologists, anesthesiologists, and emergency medicine specialists to understand exactly how oxygen loss affects the body.

And since you can't exactly recreate a breathing emergency in someone's living room, the algorithm was tested in clinical laboratories where volunteers safely mimicked emergency conditions under medical supervision.

On top of that, the system was checked against hundreds of thousands of hours of real-world data to weed out false alarms from things like sleep apnea or simply resting quietly.

Google also worked directly with emergency medical services and public health leaders to make sure the alerts slotted neatly into existing emergency response systems — so when the watch does call for help, the right people can be prepared.

To take advantage, you'll need to manually toggle it on in your device's settings.

Google is also rolling out this out to Pixel Watch 4 owners.

The Pixel Watch 5 comes with the existing Loss of Pulse Detection feature. Running similarly to breathing emergency detection, it's an advanced safety tool on the Pixel Watch designed to detect if your heart stops beating due to events like cardiac arrest or respiratory failure.

If the alert isn't dismissed, your watch or connected phone automatically calls emergency services, shares your location, and plays an automated message.

Google is also rolling out three new monthly health trend summaries covering blood pressure, sleep breathing quality, and insulin resistance if you have diabetes. These are each powered by Health Foundation Models trained on billions of minutes of sensor data and validated against clinical-grade measurements.