SimpliSafe’s new outdoor monitoring service combines AI with live agents

SimpliSafe is bringing live monitoring to its outdoor cameras, adding to the indoor live monitoring service it launched last year. The new live guard outdoor protection service lets SimpliSafe’s human agents view a live feed from its outdoor camera when the system is armed, the camera detects a person, and its AI and facial recognition algorithms determine that person is likely to be a stranger. The idea is that SimpliSafe could more proactively prevent someone from breaking in than traditional alarm systems, which typically only trigger after a home’s perimeter has been breached.

The DIY smart home security company’s new service uses a combination of on-device AI, cloud-based computer vision, and facial recognition to determine when someone is approaching the house and then decide whether they should be there or not. “We’ve built an AI pipeline that filters out people with profiles you’ve set up in the app,” SimpliSafe’s SVP of product, Hooman Shahidi, told The Verge.

If the camera identifies a face without a user-created profile associated with it, it triggers the live guard system, and the human agent then gets access to the live feed to take the necessary next steps, according to Shahidi. These can include talking with the person and warning them off, activating the siren, and dispatching emergency services. The homeowner also receives a notification in the SimpliSafe app with a recording of the interaction.

Image: Simplisafe

This is the first time the company has introduced biometric identification to its security service. Its indoor live security guard monitoring service, which works with its AI-powered indoor security camera, relies on PIR motion and person detection — there’s no facial recognition. The new outdoor protection feature uses SimpliSafe’s wireless outdoor cameras and its Video Doorbell Pro.

Amazon-owned Ring offers a similar outdoor camera monitoring service for $99 a month using its Ring Alarm Home Security System. However, that service doesn’t use facial recognition; instead, it relies on motion detection and AI-powered person alerts.

SimpliSafe has not announced pricing for live guard outdoor protection, which is part of its Early Access Program (current customers can request to join by enrolling in email updates at SimpliSafe.com). The company says it will be available to all customers later this year.

Live indoor monitoring is included in the company’s Fast Protect $29.99-a-month package, but Shahidi says it’s likely that there will be an extra charge for the outdoor service as “it will be costly to provide.” He says the company’s looking at two tiers: one that covers your property 24/7 and a less expensive option that only activates the service at night.

Image: SimpliSafe

SimpliSafe says all live and recorded videos are securely encrypted and are not accessible by monitoring agents after an event has been reviewed and closed. “All footage is encrypted in transit and at rest to ensure a bad actor can’t access any of that info,” said Shahidi. “We give the user 100 percent control relative to what the agent can see, and the agent only has access to the cameras when the system is armed, and the event has triggered it, and at no other time.”

Shahidi said SimpliSafe is one of the few smart home security companies with its own monitoring centers, and its agents are all vetted and trained by the company. In states such as Illinois, where biometric identification is protected by privacy laws, Shahidi said the agents will manually match your profiles with footage from the cameras.

This type of proactive security monitoring that helps prevent a crime or catastrophe from occurring is an interesting development in the connected home, and I expect we’ll see more of it as the full capabilities of artificial intelligence are tested. However, facial recognition — as helpful as it is (I’ve tested it on Google Nest and Eufy cameras and through Apple’s HomeKit Secure Video service) — raises a lot of privacy concerns. While SimpliSafe’s implementation is designed to cut down on false alarms and could be very useful in certain situations, for most people, it’s probably overkill.