Text to speech is ML or AI?

Question

If we use text to speech and speech to text into solutions, does it called usage of artificial intelligence? or is it called machine learning? It seems this is AI,not ML..or it is neither... what you people say ?

Answer 1

There are a few methods for text to speech, and most don't have any sort of learning or training invovled.

The most common way is to just record every possible syllable, so that the sounds can be cobstructed according to what word is being said and where in the sentence.
This is not learning, this is a database.
So it's not a AI really either.

However, machine learning text to speech does exist. Training a nueral network to make a sound depending on the location of a sentence and the word/letter based off a dataset. That is AI and machine learning.

Answer 2

Machine learning is a subtopic of artificial intelligence. So you couldn't say that something is ml but not ai.

For TTS mostly deep learning is used which (as the name implies) is a subtopic of machine learning. Deep learning uses techniques modeled after the human brain involving layers of thousands of (virtual) neurons.

Edit: Take a look a the mozilla tts project if you want to know how it looks like:

https://github.com/mozilla/TTS

Answer 3

In general, if a system needs to be trained by a large set of samples then it's an example of ML. So I would say that TTS and STT are both examples which involve ML. But as ML is just a subset of AI, they are both cases of AI as well.