Microsoft has introduced a new “Portraits” feature for Copilot users in the US, UK, and Canada. It offers 40 stylized human avatars that display emotions during voice conversations with AI. Users can choose a portrait and pair it with a voice to make interactions with the assistant more natural and comprehensible compared to text-based interaction.
The feature operates in Copilot Labs and utilizes a Microsoft Research development called VASA-1. This solution allows for the creation of facial expressions, head movements, and lip-syncing with the voice based on a single image. It does not require complex 3D modeling, simplifying the process and speeding up AI performance in real-time.
Microsoft emphasizes that the portraits are not photorealistic but intentionally stylized. This gives users a choice and helps avoid confusion with real people. All avatars have the same safety features as other Copilot capabilities, and the system clearly indicates that the interaction is with AI.
Access to “Portraits” is limited to select Copilot users aged 18 and over. Microsoft has restricted daily and session usage time for this feature and added special indicators to remind users of the AI interaction.
