Microsoft's Azure AI Speech, which was showcased at the Build conference, has a personal voice feature that replicates recorded voices. However, this feature is available on a limited basis and only for specific use cases like voice assistants. Microsoft also adds watermarks to voices generated by Azure AI Speech's personal voice and requires users to obtain full consent from people being recorded to prevent impersonation.
Key takeaways:
- Truecaller is partnering with Microsoft's Azure AI Speech to allow users to create an AI version of their voice to answer calls.
- The AI Assistant, which screens incoming calls and informs users of the reason for the call, was first introduced in 2022 and is only available in select countries.
- Microsoft's Azure AI Speech has a personal voice feature that replicates recorded voices, but it is currently available on a limited basis and only for specific use cases like voice assistants.
- Microsoft adds watermarks to voices generated by Azure AI Speech’s personal voice and has a code of conduct requiring users to get full consent of people being recorded and prohibits impersonation.