The text-to-speech isn't clear enough to distinguish.
Almost always, it's too difficult for a complete beginner to distinguish between certain words (bebes/bebo, yo/tu, etc) with the text-to-speech program currently available. I don't know if it's an issue with the text-to-speech or if I'm just not experienced enough, but I can't be the only beginner having issues.
I find the "slower" button really useful in these scenarios. I think it is useful, to not rely on understanding the pronunciation but to have to listen to the phrase again. In the majority of cases the "Yo" or "El" or "Ellas" at the start should help you to figure out the context to use "bebo/bebes" etc.
I'm constantly bothered by her pronunciation of "Yo" as "Jo", which I think is a regional dialect and not standard Spanish...at least not as spoken by any of my friends and neighbors here in California. That said, I have learned to just click the slower button before submitting my translation. It's the safest way to be sure.