My Dad Aims to Stay at Home as He Gets Older, With AI Keeping an Eye Out.

A few weeks later, driven by curiosity, I asked for the transcripts of everything Sensi was capturing in my father’s home. As I read through his private conversations, I suddenly felt like a spy, with the device serving as my silent partner in crime. I had been the one to push for its installation, but now I felt a twinge of discomfort. My father, on the other hand, had no recollection of being informed that Sensi was listening in on his conversations.

As I read his own words back to him, I prepared myself for a challenging conversation.

“So, what are your thoughts?” I asked.

There was a pause during which the sound of my own blood rushing filled my ears.

“Well,” he finally replied, sounding a bit unsettled. “It’s pretty strange that it hears words.” He seemed perplexed that anyone would consider his conversations important enough to transcribe.

“But I suppose it’s worth having,” he added, quickly changing the topic.

This story is part of The Future of Home, a collaboration between the editors of WIRED and Architectural Digest to help you understand what “home” will look like tomorrow and beyond.

After my dad’s shrug of acceptance, I began researching the technology I had placed in his home. I discovered that Sensi is among a growing array of AI devices designed for seniors: Earzz and Ally Cares monitor care home residents for signs of coughs, falls, and unusual movements, while Cherish Serenity—shaped like a chic, retro speaker—uses radar to detect if someone in a room has fallen or is slumped over. (This device can be bundled with AT&T for quick emergency response.)

Unlike Alexa, these devices don’t wait for someone to say “help.” They start recording after certain events occur, such as thuds, coughs, screams, and movements like falling from a bed. In the case of Sensi, it doesn’t even inform the senior that it’s recording, shedding light on my dad’s confusion.

Sensi’s algorithm, which allegedly draws on “1,000 years” of audio data, claims to identify anomalies in a person’s usual behavior. If you develop a new cough, spend too much time in the bathroom, or move around the house in an unusual manner, Sensi can reportedly detect that. When I inquired how the algorithm was developed, Romi Gubes, the company’s cofounder and CEO, stated only that its models are “trained on anonymized datasets” without “personally identifiable information.” She didn’t go into detail about what those datasets include or their sources.

Steve Kamau, a calm and soft-spoken coordinator at Husky Senior Care, which assists my dad with shopping and other household tasks, shares that the device sometimes operates exactly as intended. In one instance, a senior fell while attempting to reach the toilet when no caregiver was present. Sensi recorded both the sound of the fall and the man’s cries for help. Kamau contacted the client (who always carries his phone) to confirm he had fallen and subsequently called 911; help was dispatched, and the man was lifted off the floor. In another case, Kamau mentioned, a client’s cough was detected early enough to potentially prevent a more serious illness. (Sensi claims a 90 percent accuracy rate, with edge cases evaluated by a “human in the loop”; Kamau noted that the system once mistook a dropped remote for a fallen senior.)