Artificial intelligence device for learning deidentified speech signal and method therefor

This invention describes an artificial intelligence device that listens to someone speak, then changes the sound of their voice by specifically altering its pitch to make it anonymous. It then uses this anonymized speech to train a speech recognition model, which learns to understand the meaning of the original, identifiable speech. The core idea is to protect privacy while still allowing AI to learn from spoken words.

Why it matters: Filed before large language models made speech data ubiquitous and privacy a paramount concern. The need for de-identified speech for AI training is now more critical than ever.

Status: Not in forceListed as no longer active. The specific reason is not in the record we hold.
How hard to build: SpecializedHeavy software system for AI speech processing

3 AI reinvention directions for this patent

See all three

AI gives you a few directions you could take this. Pick one, and we check whether your version is different enough to patent, then write the filing.

Reinvent this with AI

Where this one came from

Every morning we go through millions of abandoned patents and pick out 15 worth your time. This was one of today’s.

This one

+10