"Masking" the artificiality of TTS voices

I have mp3 files of TTS (synthetic) voices which sound OK, but in some spots they sound artificial (strange rising intonation where it shouldn’t be etc). Is there any effect or technique that would mask this “artificiality”? E.g. some local “pitch correction” that would “even out” the odd intonation pattern or maybe some kind of a distortion that would make the voice sound like it’s coming through a phone, but at the same hide the “artificiality”?

Not that I know of. People pay extra for better voice generation to get rid of that. Even if we did get rid of voice intonation, that would just give you flat answering machine voice.

I have mp3 files

That’s actually the first problem. Audacity doesn’t edit MP3. It converts the sound to a perfect quality internal format and edits that. If you want to make a new MP3, Audacity will make the usual MP3 compression damage plus the original compression damage. If the original quality was borderline, you may have unacceptable sound damage in the new voices, even if you fixed the intonation problems.