Vivoka is introducing a new way to collect data to train AI models with their merge of Web3 and artificial intelligence – Ta-Da. Led by William Simonin, the company has just launched the private beta for the project. Public beta is expected next quarter.
Simonin explained to Decrypt that Ta-Da intends to provide a platform for diverse AI businesses, allowing them to access affordable and high-quality data. The platform will use blockchain technology to encourage users to create data, such as voice recordings, images, videos, and texts, by completing tasks like reading a sentence or recognizing an object. They will be rewarded with TADA tokens for their contributions. Built on the MultiversX blockchain, Ta-Da addresses the problem of high costs and inconsistent data quality.
Ta-Da’s model is also designed to prioritize user privacy by using only volunteer-generated data, in contrast to the practices of other companies like Meta and Amazon. The platform’s main purpose is to gather voice recordings in various languages, which can then be used to refine AI voice recognition systems. Vivoka has already worked with 100 clients and its technology is embedded in over 100,000 devices.
In Ta-Da, users will have access to various tasks with different rewards. For example, they will be paid more for tasks that require rare voices, such as someone who can speak Corsican with an English accent. Simonin concluded that the goal of Ta-Da is to provide a diverse and expansive audio dataset to understand complex languages.
