Microsoft’s Amazing Copilot Portraits: Talk Face-to-Face with AI in Real Time with 40 Realistic Avatars to Choose From

Microsoft has launched a groundbreaking Copilot Labs experiment—Copilot Portraits—which allows users in the US, UK, and Canada to engage in natural voice conversations with AI-powered, animated portraits. This new innovation is part of Microsoft’s ongoing efforts to make AI interactions more relatable, comfortable, and human-like by providing users with expressive digital faces during their chats with Copilot.

Copilot Portraits: What’s New?

Microsoft’s Copilot Portraits experiment introduces 40 stylized, animated avatars that move, speak, and respond in real time, supplementing Copilot’s voice assistant features. These digital faces are intentionally non-photorealistic—designed to offer comfortable interaction without blurring the line between AI and real humans.

• Users can select from a variety of genders, ethnicities, and nationalities, giving the experience wide visual diversity while remaining safely in the realm of “obviously a bot.”

• Portraits use Microsoft’s advanced VASA-1 AI engine, which animates expressive faces realistically from a single image, creating synchronized mouth and head movements with the user’s speech.

• Conversations with the animated avatars are live, making the dialogue more natural and engaging than static backgrounds or stylized cartoon mascots.

How to Access Copilot Portraits

Copilot Portraits is available as an experimental feature within Copilot Labs for users in the US, UK, and Canada.

• Access requires a Microsoft account and is currently limited to Copilot Pro subscribers, priced at $20 per month.

• Eligible users can navigate to Copilot Labs and find Portraits in the sidebar, where they can choose an avatar and a voice before beginning a voice chat.

• Age restrictions apply: users must be at least 18 years old, and there are clear indicators throughout the experience to remind users they are interacting with artificial intelligence.

Why Animated Portraits?

The idea for animated Copilot experiences stems from user feedback: many people are more comfortable talking to a visually expressive face than a faceless voice or plain chat window. Mustafa Suleyman, Microsoft AI CEO, stated the rollout is a direct response to those insights, seeking to make users feel “more at ease” and better connected during their AI-powered conversations.

Key motivations highlighted by Microsoft and Suleyman include:

• Providing visual cues and emotional resonance during conversations, especially valuable for brainstorming, practicing interview skills, or study sessions.

• Establishing a new standard for immersive, comfortable, and safe AI interactions.

• Experimenting with educational features like “study mode voice,” which could cater to learning or focused tasks in future releases.

Technology Behind the Feature

Copilot Portraits is underpinned by Microsoft’s VASA-1, a powerful AI system capable of real-time facial animation:

• VASA-1 generates synchronized expressions and head gestures from just one static image, requiring no complex 3D scanning or modeling.

• Rather than deep-faking photorealistic results, it keeps avatars safely stylized for both ethical and practical reasons.

• Microsoft aims to ensure all chats remain distinctly artificial, with embedded safeguards to prevent inappropriate or misleading user experiences.

Safety, Ethics, and Microsoft’s Cautious Rollout

AI companions with lifelike visual features present new risks, including over-attachment, confusion about AI’s identity, or even possible inappropriate behaviors (issues currently scrutinized at rival companies). Microsoft’s approach includes:

• Strict access and age gating (18+)

• Daily session time limits

• Persistent reminders that users are speaking to AI

• Enhanced moderation and feedback channels

• Deliberate stylization of avatars to avoid the “uncanny valley” and minimize risk of confusing AI with real people

The experiment is currently in a limited rollout to gather feedback and refine the experience before any broader release.

Early User Feedback and Future Prospects

Initial responses have been positive, with users noting smooth language switching, spontaneous storytelling, and engaging character behavior. Some observed that while avatars are charming and interactive, they maintain clear boundaries, such as not singing or fully participating in complex poems—reinforcing the intent to keep them comfortably artificial.

Microsoft has stated that further refinements are on the way, and its Copilot Labs framework is open to wider user suggestions and testing. The ultimate goal is to discover how people feel most at ease with AI, and to give every Copilot interaction a distinctly “personal” touch, while ensuring safety and transparency remain top priorities.

The Competitive Landscape

Microsoft is not alone in experimenting with visual AI avatars. X (formerly Twitter) recently introduced 3D avatars—including an “anime-inspired AI girlfriend”—for its Grokbot, while Character.AI and other startups have stumbled with moderation challenges around lifelike AI companions. By focusing on transparency, safety, and stylization, Microsoft aims to set itself apart with Copilot Portraits.

Copilot Portraits marks a significant evolution in how users can interact with AI, blending expressive visuals with Microsoft’s robust conversational technology in a safe, tightly controlled environment. As these digital companions become more natural and engaging, Microsoft is carefully shaping the ethics, safety features, and user experiences that will define the next era of human-AI interaction.

Try Microsoft Copilot Portrait feature right now.