With Siri, Apple has launched a working voice control system with text recognition, which naturally motivates Google to develop a similar product. However, Google's variant is supposed to be reminiscent of the computer from Star Trek - hence the project's name "Majel", since the original English version is spoken by the computer voice of Majel Roddenberry. The software is being developed in secret, so it is not yet known how it will work in detail. However, a first version is expected to be released this year, which does not mean that much time is left.
Whether this will enable truly natural voice output on smartphones - as in the series - is still open. Google has already taken over Phonetic Arts, a company that specializes in speech synthesis, so the plans do not seem completely out of the question.
My opinion: voice control may sound tempting, but why on smartphones first? Wouldn't it make more sense to have a working control system à la Star Trek in your own four walls? There, the data could also be transferred to servers in the cloud without any problems in order to process the voice inputs. And: you wouldn't have to disclose your appointments to everyone in the subway.