Company Tries to Make Synthesized Voices Sound More Like Us

Alexa sometimes has an eerie laugh users find disturbing, but Amazon’s digital assistant’s voice still sounds electronic—like a computer voice. A start-up called Voicery may change that. The company is working to develop custom digitalized voices that sound much more like a human’s, based on the sound of professional voice actors who record 200 hours of speech for AI to imitate.

In a story in Fast Company’s Design section, Voicery CEO Bobby Ullman said he is working with co-founder and CTO Andrew Gibiansky to develop AI-synthesized voices that mimic human emotion—at the request of companies that want “branded” voices to represent them.

“As more of our interactions with companies shift away from the visual and toward the verbal–whether thanks to Echo and Google Home or automated customer service systems–the tone, quality, and cadence of a company’s voice is becoming the new face of the brand,” writes Katharine Schwab in the article.

Voicery’s website provides a quiz which allows the reader to guess whether the voice that plays is human or a machine. Some of the voices are difficult to distinguish. To try it out, click here.

The company plans to create a library of hundreds of thousands of voices in a variety of languages.