Millions of smart home security systems are getting a bit smarter this week. A new feature called Smart Arming brings a little flexibility to Alarm.com systems’ arm and disarm settings, making you less likely to trip the alarm by mistake and more likely to actually use it.
Smart Arming is available now on systems powered by Alarm.com, a company that provides both hardware and software to hundreds of local and national home security companies. Instead of a rigid on / off schedule, Smart Arming lets you schedule arm and disarm windows, then uses information from security sensors in your home that you select to determine when to set the alarm or disarm it within those windows. It’s like the difference between a programmable thermostat that follows its schedule no matter what and a smart thermostat that can modify its behavior depending on whether anyone’s actually home.
Smart Arming doesn’t use AI or machine learning
With Smart Arming, you won’t accidentally trip the alarm if you leave on your morning run at 5:30AM instead of 6AM — the motion sensors can detect you moving toward the door and disarm the system before it wakes the whole house. And if you head out to the fire pit one evening when the alarm is usually about to arm, the system can use data from the contact sensor on the door to wait until you come back in before arming itself.
Despite being marketed as “intelligently” adjusting your alarm system, Smart Arming doesn’t use AI or machine learning. “It’s taking input from sensors throughout the home to determine when to disarm. It’s not learning your patterns, but it does understand them,” Abe Kinney, senior director of product management at Alarm.com, told The Verge. Essentially, it’s a slightly more nuanced set of rules for your alarm system.
This type of innovation is a bit unusual in the smart home security space. Many “smart” systems are only smart because you can control them with an app, and they can tie into a platform like Amazon Alexa or Apple Home. The actual security side is still largely what it’s always been: if a sensor is triggered when the alarm is armed, the system alerts a homeowner / calls the monitoring service — without many nuances beyond that. This is because, when you’re playing with people’s safety, you must know you are getting it right, and simplicity is the safest option.
Alarm.com, and other companies, do deploy machine learning features for less critical areas, such as on cameras, to identify people over other motions. Alarm.com also has a feature called Unexpected Activity alerts that use its “Insights Engine” to learn a household’s behavior patterns and send an alert if something unusual happens beyond a standard security event.
Overall, features that make alarm systems more user-friendly — more convenient to arm and less likely to trigger accidentally — are a good thing. Many people (myself included) worry about arming a system because the cat, a child, or a visitor may inadvertently trigger it, and they don’t want to deal with the whole household waking up or the police knocking on the door for a false alarm.
Features like this make alarm systems more user-friendly — more convenient to arm and less likely to trigger accidentally
While many smart security systems can be set to arm and disarm automatically based on the geolocation of your household’s phones, that only helps when you leave the house, not so much when you go to sleep at night.
Others — such as Ecobee — use facial identification to disarm automatically when someone it recognizes comes home. But the idea that your alarm system can know no one has broken in when your preschooler goes to the kitchen for an early morning snack without needing a camera in your home is even better.
The obvious downside here is if there’s potential for the system to get it wrong: someone breaks in but the alarm doesn’t go off because there had been some activity inside a few moments before. Kinney says this concern is why the feature requires time windows, so if there is something that goes bump in the middle of the night, you’ll be sure to know about it, he says.