At this year’s Microsoft Ignite conference 2023, the tech giant launched a revolutionary technology i.e. Microsoft Deepfake Creator that enables users to create photorealistic avatars that can convey customized messages—a startling development. Known as Azure AI Speech text-to-speech avatar, this tool allows users to upload photographs of themselves and add a script to produce films. It is now in public preview. Next, the technology uses a text-to-speech engine to narrate the scripted information and a model to animate the avatar.
Microsoft Deepfake Creator:- Azure AI Speech text-to-speech avatar
Microsoft claims that with just text input, this new tool makes it easier to produce training movies, product introductions, client testimonials, and more in an effective manner. Not only can the avatars speak many languages, but they can also use AI models, such as OpenAI’s GPT-3.5, to answer natural language questions in chatbot scenarios.
Although there are many possible uses for this technology, Microsoft recognizes that there are moral questions that can arise from its use. As a result, the majority of Azure subscribers will initially only be able to access prebuilt avatars; personalized avatars will only be available to those with restricted registration privileges and for certain use cases.
The introduction of this tool brings up issues that are similar to current disputes in the industry, such as the SAG-In the recent past, we have seen various bad causes of Deepfake AI like Rashmik Mandhana Indian Actress video, and many more. AFTRA strike, in which performers sought payment for the use of their likenesses created by artificial intelligence. Microsoft did not answer to questions, so it is unclear what position the corporation takes on the moral ramifications of utilizing actors’ likenesses without giving them credit or notice.
Personal Voice By Microsoft
Nevertheless, Microsoft appears to be exercising greater caution with another tool—Personal Voice—that was presented at the Ignite event in 2023. With just a one-minute speech sample, this component of Microsoft’s proprietary neural voice service can mimic a user’s voice. With Personal Voice, individuals must give their explicit agreement via a recorded statement before their voices may be synthesized. Personal Voice is intended to create personalized voice assistants and translate content into many languages.
Microsoft has imposed enough required limits on the use of Personal Voice in order to prevent possible legal problems. The capability can only be used in apps when the voice does not read user-generated or open-ended information, according to user agreement. Furthermore, the speech models’ output must not be shared or published outside of the program.
In spite of these steps, it’s still unclear how actors will be paid for their work on Personal Voice and whether Microsoft intends to use watermarking to make artificial intelligence sounds identifiable.
The industry keeps a close eye on Microsoft’s advancements in artificial intelligence, attempting to strike a balance between innovation and morality when using these state-of-the-art instruments.