That's why it's called Text-to-Speech. It's the opposite of the Speech-to-Text that they've had since Chrome 10.