Microsoft launches AI tool for photorealistic copying of faces and voices | Artificial Intelligence (AI)

Microsoft announced its latest contribution to the artificial intelligence race at its developers conference this week: software that can generate new avatars and voices or replicate a user’s existing appearance and speech – sparking concerns that it could boost the creation of deepfakes, created by AI. videos of events that didn’t happen.

Announced at Microsoft Ignite 2023, Azure AI Speech is trained with human images and allows users to enter a script that can then be “read” aloud by a photorealistic avatar created with artificial intelligence. Users can either choose a preloaded Microsoft avatar or upload images of a person whose voice and likeness they want to replicate. Microsoft said in a blog post published Wednesday that the tool could be used to create “conversational agents, virtual assistants, chatbots and more.”

The message reads: “Customers can choose a predefined or custom neural voice for their avatar. If the same person’s voice and likeness are used for both the custom neural voice and the custom text-to-speech avatar, the avatar will sound very similar to that person.

The company said the new text-to-speech software is released with various limitations and safeguards to prevent misuse. “As part of Microsoft’s commitment to responsible AI, the text-to-speech avatar is designed to protect the rights of individuals and society, enable seamless human-machine interaction, and counter the proliferation of harmful deepfakes and misleading content,” the company said. .

“Customers can upload their own video recording of the avatar talent, which the feature uses to form a synthetic video of the custom avatar speaking,” the blog says. The “avatar talent” is a human posing for the proverbial AI camera.

The announcement quickly sparked criticism that Microsoft had launched a “deepfake maker” – which would make it easier to mirror a person’s image and make them say and do things they didn’t say or do. Microsoft’s own president said in May that deepfakes were his “biggest concern” when it comes to the rise of artificial intelligence.

In a statement, the company pushed back against the criticism, saying that custom avatars are now a “limited access” tool for which customers must apply and be approved by Microsoft. Users will also have to disclose when AI was used to create a synthetic voice or avatar.

“With these safeguards in place, we help mitigate potential risks and enable customers to integrate advanced voice and speech capabilities into their AI applications seamlessly and securely,” said Sarah Bird. , from the division responsible for AI engineering at Microsoft, in a press release.

The text-to-speech avatar maker is the latest tool as big tech companies rush to capitalize on the artificial intelligence boom of recent years. After the meteoric popularity of ChatGPT – launched by Microsoft-backed OpenAI – companies like Meta and Google have launched their own artificial intelligence tools into the market.

The rise of AI has been accompanied by growing concerns about the technology’s capabilities, with OpenAI CEO Sam Altman warning Congress that it could be used for election interference and that Guarantees must be implemented.

Experts say deepfakes pose a particular danger when it comes to election interference. Microsoft launched a tool earlier this month to allow politicians and campaigns to authenticate and watermark their videos to verify their legitimacy and prevent the spread of deepfakes. Meta announced a policy this week requiring disclosure of the use of AI in political ads and prohibiting campaigns from using Meta’s own generative AI tools for ads.