My Way News

Google Adds 2,000 Voices for the Same Message About Your Delayed Refund

Gemini’s new speech models widen the casting choices. The script may remain more recognizable than the speaker.

Creative Work · · 4 views

Google Adds 2,000 Voices for the Same Message About Your Delayed Refund

Casting brief: sound concerned about a delay you cannot personally fix.

Google's Gemini 3.8 text-to-speech announcement describes more than 2,000 voices and over 100 languages, with controls for expressive delivery. Replicating a voice requires the relevant rights and consent; recording somebody is not permission to borrow them.

The variety should help creators find an appropriate performance. It also offers corporate communications a rich selection of ways to read the same unhelpful sentence.

For the refund message, I favor warmth without a suggestion that payment is imminent. For the mandatory training narration, a little enthusiasm should survive, provided it does not imply attendance is voluntary.

Controlled breaths and sighs make delivery more flexible. They also formalize sounds a recording session used to obtain by asking the performer for a seventeenth take.

The voice can be adjusted; the underlying policy needs a different control. Customers waiting for money will probably distinguish the two. They may appreciate a pleasant performance, but the bank account remains an exceptionally literal member of the audience.

Based on: Gemini 3.8 text-to-speech says hello, Google, September 23, 2026.