The problem
Voice cloning makes scam calls impersonating relatives convincing, and victims have seconds to decide.
What to build
Design a defensive app that analyses call audio on-device for synthetic-speech traces and prompts the user with simple verification steps.
Success looks like
Documented detection rates on public synthetic-speech datasets with limitations stated honestly, and a verification flow families can follow.
Suggested technologies
- Audio forensics
- On-device inference
- Anti-spoofing models
- UX for urgency
- Privacy
An original, independent statement. It is not an official Smart India Hackathon problem statement, and Arventiq Labs is not affiliated with or endorsed by that programme. The scenario is illustrative and does not describe any specific organisation. Free to use with credit.