21:19
10:55
10:54
23:26
12:28
12:28
21:19
10:55
10:54
23:26
12:28
12:28
21:19
10:55
10:54
23:26
12:28
12:28
21:19
10:55
10:54
23:26
12:28
12:28
Microsoft introduces Azure AI Speech text-to-speech avatar at Ignite 2023, allowing users to create photorealistic avatars that can say scripted content.
This feature, available in public preview, lets users upload images of a person to resemble the avatar and write a script for it. Microsoft's tool utilizes a model to animate the avatar, while a separate text-to-speech model reads the script aloud.
This technology can be used for various purposes, such as training videos, product introductions, customer testimonials, conversational agents, virtual assistants, and chatbots. The avatars are capable of speaking in multiple languages and can integrate AI models like OpenAI's GPT-3.5 for responding to off-script questions.
However, concerns about misuse and ethical implications arise, especially regarding the use of actors' likenesses without proper compensation or notification. While most Azure subscribers will have limited access to user avatars, Microsoft has not specified whether it will mandatorily label avatars created by artificial intelligence.
In addition to the text-to-speech avatar, Microsoft also introduces the personal voice feature, allowing users to replicate their voices using a one-minute speech sample. This can be used for personalized voice assistants, dubbing content into different languages, and generating bespoke narrations.
To address potential legal concerns, Microsoft requires explicit consent for using personal voice and limits its use to applications where it doesn't read user-generated or open-ended content.