Block, the parent company of Square and Cash App, has introduced a new artificial intelligence model named Dessa. The company claims this is the first AI model developed from scratch using exclusively human speech data, bypassing text-based training entirely.
Unlike traditional systems that convert audio into text for processing, Dessa analyzes speech directly in its audio format. This approach allows the model to detect nuances such as sarcasm, laughter, pauses, and emotional states in real time.
The landscape for emotionally aware voice AI has become increasingly competitive. Recent developments include the launch of Hume AI's EVI in March 2024 and the release of Moshi by Kyutai in September 2024. Additionally, the Open Empathic project by LAION has been active since late 2023.
OpenAI also continues to refine its Advanced Voice Mode for ChatGPT, which received a major update in June 2025. These improvements focus on enhancing intonation and the expression of empathy. However, the industry continues to debate whether natural-sounding speech is as critical as the model's underlying reasoning capabilities.
