Jack Dorsey’s Block launches Dessa AI voice model, but users aren’t happy

'We trained Dessa from scratch on nothing but human conversation,' the company says

|
Published October 24, 2026
Jack Dorsey’s Block launches Dessa AI voice model, but users aren’t happy

Jack Dorsey, the man behind Twitter, now known as X, has launched an AI voice model called Dessa.

The most distinguished thing about Dessa is that the model does not run on any kind of large language model, instead it is trained on speech-to-speech human conversation.

Unlike it, other AI models such as Gemini and ChatGPT convert your voice into text before replying. But this model does not use any kind of text.

As per the company Block, the model is capable of listening and speaking simultaneously while assessing and expressing emotions such as laughter, frustration, sadness through natural voice interactions.

“We trained Dessa from scratch on nothing but human conversation. Sounds! No text, no pretrained models. It learned to hold a real-time conversation simply by listening,” the company said.

Even the model can independently decide its reactions like what to say and when to respond and what tone it should adopt during the conversations. “Talking with it feels like talking with a person...full of emotion, nuance, pauses, and a natural cadence,” it said.

Speaking about reacting to jokes and humour, Dessa can break into laughter mid-sentence and adjust its volume when asked.

The launch of the model took the internet by storm, drawing mixed response from the users. Some really loved this technology based on human interaction.

Some users raise their concerns. According to one user, the model will help in proliferating scam calls, stating, “Get ready for a new era of scam calls! Now if someone calls you’ll not be able to distinguish if it’s AI or a human, it’s pretty insane.”

Another user wrote, “That's kinda cool, but many people will get attached to it and isolate themselves from real humans even more. Many will get scammed on the phone, especially grandpas and grandmas.”

“If boomers were cooked by AI, this is just gonna make it worse. Just imagine endless spam and scams called with perfect intonation and mannerisms in the voice of someone you know. The crazy part is how natural the AI voice sounds,” the third one wrote.

Some called it “scary” and “way more dramatic” and sound like "illiterate teenager.”

Aqsa Qaddus Tahir
Aqsa Qaddus Tahir is a reporter dedicated to science coverage, exploring breakthroughs, emerging research, and innovation. Her work centres on making scientific developments understandable and relevant, presenting well-researched stories that connect complex ideas with everyday life in a clear, engaging, and informative manner.
Share this story: