I dislike how LingQ are using AI-generated audio for a lot of the Mini Stories now. I mean I can understand them using them for newly released languages, in order to expedite them in to the LingQ community ecosystem, however there was previously a perfect set of human-read Standard Arabic Mini Storie audios which were switched for AI audios. Latin American Spanish also uses AI audio, although I can’t vouch for how they previously sounded. Anyway, my point is that they don’t sound as organic or authentic as human-read audio files, which are more realistic.
This is news to me…on the Latin American Spanish. I haven’t listened to it in a year so maybe it’s changed. Did they mention this? If not, what is your evidence that it is AI audio? That would be a huge departure from their statements in the past (or at least as I’ve interpreted them).
I haven’t used content uploaded to LingQ in a long time. The reason is that I don’t like transcripts that are AI-generated and full of mistakes, or audio that is AI-generated, or old content that has an outdated language in which I’m not interested.
I can’t filter out all that content in the very beginning, and it’s faster to search and find my own content.
I don’t know about those official mini-stories with AI audio. They shouldn’t be there. IMHO.
Are you sure you didn’t run into some bug? I checked a bunch of LA Spanish mini-stories and none of them seemed to have AI audio.
My evidence is that I can tell the difference in what I’m hearing between AI and human speakers, just as one can differentiate between male and female, adult and child or even black and white speakers. It just has a robotic monotone flow and pace.
I mean it’s possible that it’s somehow reading the Mini Stories from a text-to-speech format, but that wouldn’t explain why it’s only the Standard Arabic and Latin American Spanish Mini Stories that’s doing this and not the Egyptian and Levantine Arabic or European Spanish Mini Stories that are being read in a computer voice.
I know they can do voice cloning, but I recognize the Latin American voice as the lady who does the fable cottage tales, which have been around before all the AI hype… It doesn’t sounds like AI to me, but maybe I just can’t tell as well.
I did go back and listen to some of the Latin Spanish Mini Stories and it hasn’t changed for the past few years or more. The recordings came out well before some of the fancy AI voices and even the best of these sound artificial. The Latin Spanish voice does not sound this way at all. There is different emphasis on words within a sentence that you just don’t find it AI voices in my experience. I’m quite certain it’s a human…and again goes back to what they’ve stated in the past and what they accept for these. They aren’t looking for AI voices for the recordings.
The Arabic ones…I have no idea. Not sure if I’d be able to tell the difference as I don’t know anything about that language.
On the “glitch” idea…what device are you listening to these on? Have you tried other devices? Have you tried the web?
BTW if you are clicking on individual words, THAT is TTS and will most certainly sound artificial. Also anything you import that doesn’t have audio will have the TTS voices.
We don’t use A.I voice for the Mini Stories, all stories were recorded by native speakers.