With nearly half of American adults now using voice-controlled digital assistants, AI-powered voice management has become a major trend in smart home security technology, Pew Research reports. Voice assistant integration was the top highlight in smart homes at this year’s Consumer Electronics show, BI Intelligence says. Voice-controlled security camera systems were prominently on display at the event, with several major providers introducing systems that integrate with popular digital assistants such as Amazon Alexa and Google Assistant.
As digital assistant services become integrated with smart home technology, the artificial intelligence that underlies voice control is becoming a key component in smart home security. Here’s a look at three ways AI is helping make smart homes safer.
Making Security Surveillance Smarter
One home security issue AI is helping solve is the challenge of analyzing surveillance footage efficiently. Smart home security cameras that run 24 hours a day, seven days a week on one or more monitors may generate an enormous amount of footage, creating the practical problem of how to sift through all this data to isolate relevant clips. For instance, home security camera systems smartphone integration now enables homeowners to view security monitors remotely from any location at any time. However, while this is extraordinarily convenient, merely being able to view footage does not automatically single out which footage represents a potential intruder or an actual break-in.
Artificial intelligence helps solve this problem by automatically reviewing and analyzing surveillance video in order to isolate relevant footage. AI-powered machine learning can automatically identify images that represent a trespasser on premises, triggering an instant alert sent to homeowners or security monitoring services to facilitate a faster response. AI machine learning can also reduce the rate of false alarms by distinguishing images such as birds or weather events from actual intruders, increasing surveillance efficiency. AI further allows reviewers of security footage to automatically scan through hours of surveillance in order to quickly identify scenes depicting suspects or vehicles of interest.
Providing Voice Authentication
Another smart home security challenge AI is helping solve is voice impersonation. While voice control is convenient for homeowners, it also creates a potential opportunity for voice impersonators to hack smart home systems. For instance, one woman found herself the recipient of unauthorized Amazon orders until she discovered her parrot was placing orders over Alexa while she was out. Children or relatives who sound similar to device owners may also be able to fool voice-controlled systems. Hackers who gain access to homeowner voice samples through webcams or social media may be able to use voice imitation software to impersonate homeowners and access their smart home systems.
To prevent this, smart home technology providers have used AI to develop voice authentication software that can distinguish genuine device users from imitators. While this technology is not yet perfected, it continues to improve, and high demand will drive innovation. For instance, ID R&D has recently developed voice authentication software that uses 100 distinct voice characteristics to authenticate individual users, reducing the amount of sample required to authenticate a voice from the industry average of 7 to 10 seconds down to 1 to 2 seconds, with an equal error rate of 0.3 percent. Global demand for this type of voice authentication biometrics is expected to increase at a compound annual growth rate of 19 percent between 2018 and 2023, Research N Reports projects.
Automatically Intercepting Hackers
The risk of voice impersonation is only one facet of a broader challenge posed by smart home hacking. Methods that hackers use to penetrate home computer networks can also be used to breach smart home networks. For instance, a common method of compromising home networks is hacking devices that still have their default passwords, such as routers and printers. Hackers can use this technique to compromise devices such as smart home thermostats in as little as 30 minutes, Ben-Gurion University research has found.
To prevent this, security providers have harnessed the power of artificial intelligence to automatically detect and defuse intrusion attempts. For instance, security provider Symantec recently introduced the Norton Core Wi-Fi router, which uses artificial intelligence to protect smart home devices and networks. The router draws from an AI machine learning analysis of 300 million endpoint users in order to identify attack patterns instantly and automatically stop them.
By making security surveillance smarter, biometrically authenticating user’s voice and automatically intercepting intrusion attempts, artificial intelligence is already helping make smart homes safer. As AI sophistication continues to advance, artificial intelligence innovations will continue to improve smart home networks’ ability to stop physical and digital intruders, allowing homeowners to rest securely as they enjoy the advantages of home automation.