For people without IT background, I can recommend Mozilla common voice. They plan to release an advanced AI model for text-to-speech or speech-to-text conversion, e.g. for an offline, open source alternative to Amazon Echo, Siri etc.
To train the model they need at least 10000 hours of speech samples per language. So you can donate your voice by reading aloud small snippets, checking already recorded samples or making up new sentences.