Jack Dorsey's Block unveils voice AI that mimics human emotion
Block releases a voice AI that sounds like a person
Jack Dorsey's payments company Block released a new artificial intelligence model called Dessa. It processes human speech directly instead of using text to make its responses. Block says the model can listen while speaking and reacts to interruptions, laughter, and changes in tone during a conversation.
The company shared demonstrations showing Dessa laughing mid-sentence and reacting with what appears to be genuine emotion. Dorsey described the experience on X as feeling like talking to a person, full of nuance, pauses, and natural rhythm.
Reactions from experts and the public
- Some users on X praised the natural quality of the AI voice. Alexander Hoff stated he had never heard anything artificial that sounded so human.
- Binance founder CZ Zhao suggested the technology could help with children's education by making AI voices more natural for learning.
- Filmmaker Justine Bateman questioned the need for AI to sound human and raised concerns that such realism could fuel phone scams and deception.
How the technology works
Block trained Dessa from scratch using a single Nvidia GB200 NVL72 computing system. Unlike other systems that convert speech to text first, this model processes audio directly. It is compact enough to run locally on a MacBook Pro with an M4 Max chip or better.
Current status of the model
Dessa is currently a research preview and not a public product. Block has not disclosed the model's size, benchmarks, or a release date. The initial release is going to a small group of AI safety researchers.
Sources
Newisty Editorial Team covers technology, cryptocurrency, digital products, online platforms, developer tools and the wider digital economy. Our content is researched from official sources, company announcements, public documentation, market data and other primary or reputable sources. Articles are reviewed and edited before publication for clarity, accuracy and useful context.