In our essay, “Naming the Fear,” we named the four risks we had to face building Edgecare: cameras in the home, sensitive data, sharing, and choice. Here we want to set out how we’ve chosen to address them.
Edgecare is still in development. These are the principles we are building it to, and we’re publishing them so you can hold us to them. If one ever changes, we’ll say so on this page, with the date.
No video is stored. Edgecare captures brief snippets through the day, not a continuous recording. Each snippet is analyzed and then deleted. What we keep are the measurements derived from it, such as activity and mobility, and any alerts. Those are encrypted.
The person living with Edgecare, at home or in a care facility, chooses who receives alerts. Usually that’s family, plus their doctor if they choose to include them. Nobody gets access by default, including family.
The people you choose hear from us only when there’s something they need to know. They get a short summary of the alert and why it was triggered, with a suggestion to see a doctor. They don’t get video, and they don’t get detailed measurements.
An alert is information for the people you’ve chosen. It doesn’t make decisions for anyone.
We don’t use anyone’s data to improve our models unless the person living in the home explicitly opts in.
Edgecare staff don’t see your measurements or any raw data.
If you choose to take part in testing, the people working on the product with you may see your measurements. We’ll ask first, and you can stop at any time.
We never use AI to generate people: no AI-made seniors, families or caregivers, and no AI-made footage of Edgecare at work. The people and the product you see in our videos and images are real. We don’t use synthetic or cloned voices.
We allow AI-generated content in two limited ways. The first is a background, used to hide a cluttered or private setting; it carries an on-screen label saying the background is AI-generated, for as long as it’s visible. The second is an illustration that is clearly not a photograph, such as a cover image for a blog post; we label it as AI-generated wherever it appears.
Where the analysis runs. We’re experimenting with different options, and we haven’t finalized that. What we’ve committed to is what’s kept: no stored video, only the measurements, and you decide who gets alerted.
If something here is unclear, or there’s a risk we haven’t named, write to us at info@edgecare.ai. We read every message.