Meta announced Hologram for calls from Meta Ray-Ban Display at Connect on September 23, 2026. After a short capture of the user's face in the Meta AI app, the feature creates a photorealistic digital likeness. Meta says early access will begin on WhatsApp for users of those glasses in the United States later this fall.
The key technical detail is how that likeness moves. Meta says this version is driven entirely by the caller's voice: the model infers expressions from what the person says and how they say it. It is therefore a generated representation, not a live camera measurement of every facial expression.
What changes in a conversation
A wearable call can leave the user's hands free while giving the other person a recognizable face to look at. That is a plausible convenience for informal conversations during everyday activities. It also changes the meaning of the visual cues the recipient sees.
A smile created from vocal tone may look natural without matching the speaker's actual expression. Silence, irony and mixed emotions are useful cases to test because speech and facial behavior do not always communicate the same thing. Realistic appearance alone cannot establish accurate emotional representation.
NEURA analysis: make the representation clear
A good call interface should tell both people when an avatar is active and make switching modes straightforward. That would help avoid confusion between generated facial motion and direct video, especially when a screenshot or recording is viewed later outside the original conversation.
The announcement should also not be interpreted as a promise that an autonomous copy can attend a call on a person's behalf. Generating a face from someone's speech is a different capability from generating that person's answers or making decisions for them.
The next useful evidence will come from ordinary conversations: how expressions hold up over time, how interruptions are handled and whether callers understand what is being synthesized. Hologram's appeal lies in making remote presence easier, but the interface needs to preserve a clear understanding of who is speaking and what the picture represents.
Cover: AI-generated conceptual illustration, not a screenshot of Meta Hologram.
