The AI Helping You Keep an Eye on Mom ... While Guarding Her Privacy

The AI Helping You Keep an Eye on Mom ... While Guarding Her Privacy

Why you should care

Because caregiving is hard. Artificial intelligence can make it a little easier. 

As you approach middle age, a new reality is likely to set in: caring for an aging adult. Becoming a caregiver — whether it’s for a parent, older relative or spouse — impacts a quarter of Americans between the ages of 45 and 64, according to Pew Research Center. But many in this demographic also have jobs, children and myriad other responsibilities. It’s a difficult balance, but artificial intelligence (AI) is being designed to make it easier — and protect privacy too.

A new home security system, Cherry Home, created by startup Cherry Labs, aims to alleviate some of this stress via a real-time video stream of an elderly person in their home. The system, backed by GSR Ventures, uses vision sensors and microphones to detect and track the person’s movements, providing caregivers with round-the-clock livestream monitoring. But it’s not only about having the ability to check in anytime. The system can also detect and send alerts when something’s not quite right. It’s the first AI smart home monitoring system that detects movement and patterns, rather than faces alone, making it easier to care for elders remotely. livestream monitoring. But it’s not only about having the ability to check in anytime. The system can also detect and send alerts when something’s not quite right. It’s the first AI smart home monitoring system that detects movement and patterns, rather than faces alone, making it easier to care for elders remotely.

But this 24/7 surveillance system is, says Cherry Labs, less creepy than it sounds. The video, watched on a computer or phone, doesn’t actually show the person being monitored — the camera identifies people by their face, length of their limbs, clothing color and their gait or posture. On screen, all viewers see is a stick figure, or virtual “skeleton,” which maintains a level of privacy for the person needing assistance.video, watched on a computer or phone, doesn’t actually show the person being monitored — the camera identifies people by their face, length of their limbs, clothing color and their gait or posture. On screen, all viewers see is a stick figure, or virtual “skeleton,” which maintains a level of privacy for the person needing assistance.

App anomaly

When a change in behavior occurs, the caregiver is sent an alert.

Each year, one in four Americans aged 65 and older has a fall — they’re the leading cause of fatal injury in older adults, according to the National Council on Aging. Cherry Home’s algorithm can detect a situational anomaly like a fall and sends an alert to the caregiver via text message, Facebook Messenger or WhatsApp. The system can also detect changes in activity patterns like sleep disruptions or more frequent trips to the bathroom, which can be indicative of larger health concerns.

If you live with the person you’re caring for, you’ll notice these changes right away, says Cherry Labs president and co-founder Nick Davidov. “But if you’re not there or you’re in a different city, it’s just impossible.” The larger goal of the company is to let elders stay at home, something that 75 percent of adults 50 and over say they want to do, according to AARP. “We want everyone to live safely and age in place,” Davidov says.

To equip a two-bedroom home, Cherry Home costs around $1,800, with one sensor per room. Then there’s the video subscription service, which is $30 per sensor, per month. That’s far less than the cost of the average nursing home or assisted living. Nationally, assisted living facilities average about $43,536 per year, and nursing homes cost between $82,128 and $92,376 per year.

Right now, Cherry Home is only sold through elder care agencies and health care providers, but Davidov says they may eventually move to a direct-to-consumer model. elder care agencies and health care providers, but Davidov says they may eventually move to a direct-to-consumer model.

Unusual behavior

Unusual behavior is detected by the system’s algorithm to alert the caregiver of potential danger.

Still, the idea of being monitored 24/7 is a little disconcerting, even if you need daily assistance. “The system clearly communicates outside the home, so someone could take advantage of that,” says Jacob Furst, a professor of computer security at DePaul University and an expert on AI, security and privacy issues. One possible scenario, says Furst, is that a criminal is watching the feed, sees the subject go to bed, and decides to break in. They could then hack the true feed to look as if nothing was happening. “Anytime you have information going out over the internet, you face the risk that someone can read the information and modify it,” Furst says.

The risk is always there, but Cherry Labs claims privacy is a priority. All video on the platform is locally processed and nothing is stored in the cloud to make it more secure than other smart home systems. Similar devices, like Hive or Nest, use very little AI and logic in their video monitoring, says Davidov, and can’t learn normal behaviors to distinguish from anomalies. Another system, SafelyYou, also uses AI to monitor seniors but only in care facilities, not in the home.

In the future, Cherry Home aims to provide more sophisticated analyses specifically for elder care — like a stroke detection feature — and offers other smart home features that use AI-predictive models to anticipate your needs before you say them out loud. “We want to be a fully integrated system,” says Davidov. Alexa and Siri, for example, are just voice interfaces — in order to control them, you have to say something. Now, imagine it’s dark and you want to read a book. You open the book and a light turns on automatically. “We want our system to think about what’s needed and then just do it without you having to ask,” Davidov says.

OZYGood Sh*t

If you’d want to drink it, eat it, wear it, ride it, drive it; if it’d be cool to see, listen to or do, we’re writing about it.