OpenAI has created a new product that can mimic human voices, but the company isn't ready to release it just yet. On Friday, OpenAI gave a first look at its 'voice engine,' that can create an A.I. version of a person's voice using just a 15-second sample of them speaking.