Jack Dorsey unveils realistic AI model for speech generation

Jack Dorsey presented an experimental AI model called Dessa, capable of generating speech with a high degree of realism. The system was trained on real dialogues without using text transcripts.

Key features of the model

• Instant reaction to the interlocutor's speech
• Ability to interrupt and intervene in conversation
• Natural dialogue flow without delays

Unlike many similar systems, Dessa mimics human behavior, including the habit of reacting instantly to what is said. However, developers do not plan to release the model to the public due to high fraud risks.