Dear Google Gemini Team,
First of all, I want to say thank you for supporting the Amharic language in your AI system. It means a lot to us, the Ethiopian community and all Amharic speakers around the world, to finally see our language included and respected in such advanced technology. I’ve seen a lot of great improvements in Amharic text understanding, generation, and interaction — and I appreciate that deeply.
However, I’m reaching out now to express a serious concern and frustration with the Text-to-Speech (TTS) playback feature — the speaker icon that reads the text aloud after I click it. In previous versions, this feature worked very well for Amharic. It used to read the sentences out clearly, correctly, and in a natural tone. I used it all the time — not only because it was convenient, but because it helped a lot of people who prefer or need to listen to Amharic instead of reading it.
But unfortunately, after a few recent updates, this feature has completely changed — and not for the better. The Amharic playback has become confusing, broken, and sometimes totally incorrect. Words are no longer pronounced properly, tones are off, and the sentences come out jumbled. What used to be a reliable voice now sounds robotic or even incomprehensible in some cases. This is not just a minor bug — it has made the feature unusable for me and for many others I’ve talked to.
I’ve already submitted feedback before about this issue, but so far I haven’t received a response, and the problem still hasn’t been fixed. So I am reaching out again, hoping this time it will get proper attention.
Please understand: this feature is more than just a small tool. It’s an accessibility feature. There are many users — including elderly people, people with low vision, people learning to read Amharic, or just everyday users who multitask — who benefit from being able to hear the Amharic text read back to them. When it worked well, it made a real difference in how we used your app. Losing that is a big step backward for us.
We deeply appreciate the fact that you’re working to support underrepresented languages like Amharic. But that support needs to be consistent — not just in recognition, but in functionality. I kindly and urgently request that you investigate and restore the previous, accurate Amharic TTS voice. If the older version of the voice model is still available, please consider rolling it back or offering an option to choose it.
I know you receive many requests every day, but please don’t overlook this one. This issue affects a large community, and we are simply asking to have back a feature that used to work perfectly.
Thank you for your time, your commitment, and your ongoing support for Amharic users.
Respectfully,