This post is also available in:
עברית (Hebrew)
Ever wanted to have lunch with your favorite artist? Maybe ask for some advice on your startup from Elon Musk? Now, a San Francisco-based company by the name of Tavus can make this a reality with their new product, Conversational Video Interface (CVI).
The product, a lifelike video avatar that looks and sounds like a real human, aims to cater to the way humans naturally interact with each other to create an effective AI assistant far superior to chatbots. This AI replica can emulate a specific person’s appearance, expertise, and personality, and uses the power of AI to generate a real-time conversation in response to visual and voice information it receives from the user’s video input. The company claims that these digital twins can act almost exactly like the humans they represent, which is something that can significantly improve the lives of of many people by not being confined to human limitations of space, time, or memory.
The company says that this virtual twin technology allows developers to create their own interactive video experience using the set of building blocks that the company provides. The new platform provides an end-to-end developer solution that is easy to deploy with minimal setup required. It includes features such as rolling vision and a versatile conversational pipeline designed for various applications. Users can also integrate it with models like ChatGPT-4o, Claude, or other large language models (LLMs) for added flexibility. Additionally, the platform is capable of handling high traffic volumes efficiently.
On the safety front, the company has implemented several safety measures to ensure that users can only create digital twins of themselves. It also features SOC 2 compliance, automated content moderation, and anti-hallucination quality checks, According to Interesting Engineering. Despite these measures, concerns about the safety and potential misuse of generative AI persist. As AI technology rapidly advances, it will be crucial to remain vigilant and exercise due diligence to prevent online scams, unauthorized impersonation, and other risks associated with new AI products.
