Ethics
CareBestie: Building Ethical AI That People Can Trust
Jul 30, 2025
by
CareBestie
At CareBestie, we build voice based AI helpers that support a wide range of patients, many of whom are older adults and Home health teams who care for them. Our AI makes phone calls to check in after doctor visits, provide medication reminders, and listen for early signs that something might be wrong, like confusion or emotional distress. These touchpoints matter, which is why our AI must be reliable, respectful, and safe for everyone involved.
We don’t just want to talk about being ethical, we design our technology to actually follow those values in real life. That means keeping people’s private info safe, being fair to everyone, and making sure humans stay in control. Here’s how we do it.
What Makes Our AI Safe and Trustworthy
Strong protection from the start. Every call and message is locked down using powerful encryption (like a secret code) to keep personal health info safe. We follow all the important health privacy rules, including HIPAA.
Your data stays in the U.S. All personal data is stored on secure servers in the United States that meet high security standards.
Everything is tracked. We keep a secure record of every interaction what the AI said, how it responded, and if it needed help from a human. These records help us prove we’re doing the right thing and fix problems fast if needed.
Temporary special access only when needed. If someone needs to look at sensitive information for an urgent reason, they can but only for a short time and only with permission. We keep full logs of when and why it happens.
Clear decisions you can understand. We show doctors and nurses why the AI made a certain choice. For example, if the AI thinks someone might be sad, it explains which words or voice cues made it think that.
You’re in control of your data. After a call, you can log into your account to review, download, or delete your recordings within 72 hours. No need to send an email or fill out a form it’s simple and quick.
Fair for everyone. We make sure the AI treats people equally. We test to see if the AI gives different answers for things like gender or race and if it does, we fix it. We run these fairness checks every month.
Tested before going live. New AI models don’t go live right away. We run them quietly in the background for two weeks to see how they perform in real situations. This helps us catch problems before they affect anyone.
Humans always stay involved. If something serious comes up like signs of depression or confusion the AI alerts a nurse right away. A real person reviews the situation and decides what happens next.
How We Keep Ourselves Accountable
Step | What We Check | Who Does It | How Often |
|---|---|---|---|
Getting Data | Is it clean, fair, and private | ML Team | Each dataset |
Training AI | Is it accurate and fair to all types of people | Ethics and Product Team | Every time |
Putting It Online | Do alerts and logs work right | QA and Engineering | Twice a month |
After Launch | Are things still working fairly | Ethics and Product Team | Monthly and as needed |
We also have an Ethics Council that meets every few months to check our systems, talk about tough decisions, and make updates.
Why We Take This Seriously
We’ve built healthcare tech before, and we know how important it is to earn trust. AI in healthcare isn’t just about smart software it’s about real people’s lives.
When you get a CareBestie call, you’re talking to a system that has been tested and watched closely:
Calls are safe and private from start to finish
Doctors see why the AI made a decision no mystery scores
Important issues go straight to a real nurse
The design is simple and respectful for all users
We believe ethics shouldn’t just be a buzzword. It should be part of the product. You should be able to see it, test it, and count on it.
That’s how we build AI that belongs in healthcare one respectful, reliable call at a time.
