Inspiration

Every family with aging parents knows the same fear: What if they pick up the wrong call? Scammers no longer sound suspicious; they sound like your grandson, your doctor, your bank. With AI voice cloning, a single phone call can empty a lifetime of savings. Most products are for spam not scam, which is where our product comes into the picture.

We built ElderGuard AI because our elders deserve better. They raised us and protected us. Now it’s our turn to build technology that protects them.

What it does

ElderGuard AI detects AI-generated or deepfake voices in seconds, flags suspicious calls and displays a simple warning, notifies trusted family members instantly, guides the elder through safe verification questions, blocks and reports known scam patterns, and works using a simple, touch-free phone interface that seniors can actually use.

How we built it

React + Tailwind frontend for instant navigation and UI. Pre-recorded audio samples to simulate real vs deepfaked voices. Hugging Face model for voice/deepfake detection (mocked inference for speed). Mock APIs in Postman to simulate external services. Local storage to mimic onboarding and user preferences. Browser notifications for the family alert system

Challenges we ran into

Seniors need extremely simple UI, so every screen needed good design, getting realistic scam audio samples that both sound real and sound wrong, ensuring the product felt humane to the elderly user.

Accomplishments that we're proud of

Created a product flow that feels believable and useful. Delivered a prototype that can actually be launched in the real world.

What we learned

Real-time AI safety features must be frictionless, or they will never be used. Building for seniors requires radical simplicity. Protecting vulnerable communities requires empathy, not just engineering

What's next for ElderGuard AI

Use Skyflow’s vault to protect senior identity and health/financial data. Expand to WhatsApp, SMS, and email scam detection. Build a model specifically trained on scammer behavior patterns. Build a family dashboard with alerts, call logs, and intervention features.

Built With

Share this project:

Updates