Artificial intelligence is becoming increasingly humanlike, but as the technology advances, questions about whose humanity it represents are becoming just as important as what can it do.
On Nov. 9, Nuance Labs introduced Emma, a new AI persona designed to engage users in real-time, face-to-face conversations. Unlike traditional chatbots that primarily rely on text or voice, Emma can respond through facial expressions, gestures, and conversational cues intended to mimic natural human interaction, according to Nuance Labs.
Emma is also presented as a Black woman, a creative decision that raises broader questions about representation in the technology industry, particularly when considering who develops AI systems that reproduce human identities.
Emma Is A New Approach To Humanlike AI
Nuance Labs says that Emma was developed over 18 months using Nuplex 1.0, the company’s audiovisual AI model. The technology allows the AI to listen, speak, and respond visually during conversations rather than waiting for a person to finish speaking before responding.
Emma has a distinct personality, including the ability to express opinions, demonstrate curiosity, and disagree with users. The persona’s facial expressions and mannerisms are also designed to make conversations feel more natural.
The company states that Emma’s face and voice were licensed from a real actress who collaborated with its team to capture her expressions, reactions, and speech patterns.
However, introducing a Black woman as the face of this emerging technology also invites questions about the people responsible for creating it.
Who Is Behind Emma’s Identity?
A YouTube video featuring members of the Nuance Labs development team does not visibly identify any Black team members, particularly Black women. While the video alone cannot establish the racial composition of the entire company or everyone involved in Emma’s development, the apparent absence raises questions about representation in the creative process.
What role, if any, did Black women play in shaping Emma’s personality, expressions, and conversational behaviors? Beyond licensing an actress’s likeness, how are the cultural experiences and perspectives associated with the identities represented by AI incorporated into development?
The company’s website also features a section examining how AI generates emotional expressions, including joy, sadness, and anger. The comparison examples shown in that section do not appear to feature Black subjects, even though Emma is presented as a Black woman.
Those examples do not establish which racial groups were represented in the model’s training data. Still, they invite another question: As developers work to teach AI how to express human emotions, whose faces and expressions are being used to define what those emotions should look like?
Who Gets To Define Identity In The Age Of AI?
As artificial intelligence becomes increasingly humanlike, questions about identity, representation, and creative ownership extend beyond technological innovation.
Historically, Blackface involved the imitation and performance of Black identity, often by non-Black individuals, reinforcing racial stereotypes while commodifying Black expression. Could developing Black AI personas without meaningful Black participation raise similar concerns in a digital context? Where is the line between authentic representation and the commercialization of racial identity?
As Nuance Labs looks to expand its technology through additional personas and developer tools, these questions become increasingly relevant. Emma’s introduction highlights more than the growing sophistication of conversational AI. It raises a broader question about the future of representation in technology: With the increased capability of AI to reflect human identity, will the people whose identities it reproduces have an equally meaningful role in shaping it?