Google has announced that its DeepMind artificial intelligence (AI) unit has developed a new machine-generated speech engine that is 50 percent better than existing technology. Called WaveNet, the new AI speech engine has learned how to create the same sorts of sound waves that a human voice creates. It sounds much more natural than computer-generated speech created by stringing together recorded human voices.
However, WaveNet does have a big drawback: it requires a huge amount of computational power, which makes it impractical for commercial uses. Still, it’s a noteworthy step forward for artificial intelligence.