Microsoft has launched software capable of creating photorealistic avatars and voices, which has sparked fears over potential misuse for deepfake production. The post from Microsoft Ignite 2023 explains the software, Azure AI Speech, is designed with safeguards to prevent misuse, but critics suggest it could enable a more accessible creation of deepfakes.
Azure AI Speech: Creating lifelike Avatars
Microsoft's latest AI offering, Azure AI Speech, is a breakthrough technology that can create photorealistic avatars and replicate human voices. Announced at the Microsoft Ignite 2023 conference, this software allows users to 'train' it with human images and input scripts. The AI can then generate an avatar that can read the script aloud, creating a lifelike representation of the user. This is a significant step forward in AI technology and opens up new possibilities for virtual assistants, chatbots, and other tools.
Safeguards against misuse
Microsoft is aware of the potential risks associated with such high-level technology and has taken steps to ensure it’s used responsibly. They’ve put limitations and safeguards in place on their new text-to-speech software, which should help prevent misuse. One such safeguard allows customers to upload their own video recording of 'avatar talent'. The software then uses this input to create a synthetic video of the avatar speaking, maintaining control over the likeness and voice that the software recreates.
Criticism and Microsoft's response
Despite Microsoft's precautions, critics argue that the company's new software essentially acts as a 'deepfakes creator.' They suggest it could be used to replicate and manipulate a person's likeness without their approval, leading to potential harm. However, Microsoft has responded to these concerns by emphasizing its safeguards. They've stated that the customized avatars produced by the software are a 'limited access' tool that users must apply and be approved for. This adds an additional layer of control over who can use this technology and for what purposes.
Microsoft, along with other major tech firms like Meta and Google, are capitalizing on the recent surge in AI technology. But with the rapid development and deployment of these tools, concerns about their capabilities and associated risks have also risen. Microsoft has taken steps to mitigate these concerns, including the recent launch of a tool that allows politicians to authenticate and watermark their videos. This helps verify the legitimacy of their messaging and prevents the spread of misleading deepfake videos.