Can audiences connect to synthetic characters?

Much of the conversation around synthetic performers still begins with the face. Is it convincing? Is the skin realistic? Can an audience tell whether the person is generated? Those questions matter technically, but creatively they may already be the least interesting part of what is happening.

The more difficult question is not whether a synthetic character can look real for thirty seconds. It is whether an audience can care about that character for thirty days, six months or a year. That changes the challenge completely.

A convincing image can create attention. A relationship requires memory, consistency, vulnerability, contradiction and time.

A convincing image can create attention. A relationship requires memory, consistency, vulnerability, contradiction and time.  Human performers do not become meaningful to audiences because every frame is visually perfect. They become meaningful because viewers learn how they speak, what they avoid, what frightens them, how they change and what remains recognisably theirs. We build an internal model of a person. Synthetic characters will be judged by the same emotional standard.

In practice, continuity is far harder than realism.

This is the idea I have been exploring while creating Victoria D., an openly synthetic protagonist whose fictional life is designed to unfold publicly over 365 days. The experiment is not based on hiding the technology. The audience is meant to know what Victoria is. 

The interesting tension comes from asking whether that knowledge prevents emotional attachment, or whether continuity can become more powerful than the awareness of artificiality. In practice, continuity is far harder than realism.

Above: Victoria D.’s story kicked off 13 September and unfolds through 52 weekly challenges. Her mission is to access the inheritance left for her by her wealthy late husband. 

If a character appears in hundreds of short episodes, every small decision accumulates. Her visual identity has to remain coherent, but so do her memories, emotional reactions, relationships and history. If she behaves as a different person whenever the production method changes, the illusion collapses, no matter how beautiful the image is.

Brands have spent years working with influencers because influence is built on familiarity and trust rather than a single exposure. 

For me, that is where the role of human authorship becomes essential. AI can be a production medium, but a long-term character still needs direction, restraint and editorial judgment. Someone has to decide what she remembers, what she misunderstands, when she changes her mind and which parts of her personality should resist change. The audience also changes the equation.

Social platforms have trained viewers not only to watch but to respond, speculate and intervene. When a fictional character exists in the same feed as friends, creators, brands and news, the relationship can become unusually intimate. A thirty-second episode may be small, but if it arrives repeatedly and at the same rhythm as ordinary life, it can create a form of narrative companionship. That possibility is especially relevant to advertising.

A synthetic ambassador can theoretically be placed anywhere, styled in any way and made to deliver any message.

Brands have spent years working with influencers because influence is built on familiarity and trust rather than a single exposure. Synthetic characters could offer a new version of that relationship, but only if brands resist treating them as infinitely controllable digital mannequins.

Perfect control is tempting. A synthetic ambassador can theoretically be placed anywhere, styled in any way and made to deliver any message. But if every partnership bends the character instantly, the audience learns that there is no personhood to believe in. The character becomes a format, not a relationship. The most interesting commercial question, then, may be whether a synthetic performer can say no.

Above: Each episode sees her open a sealed envelope from her much-mourned beloved that will yield vital clues. 

A believable character should have boundaries. Certain products should not fit her. Certain tones should feel wrong. A campaign should sometimes adapt to the character rather than the character adapting completely to the campaign. That friction may sound inefficient, but it is also what makes human endorsements credible. Consistency creates value precisely because it limits what can happen. Transparency matters for the same reason.

We already cry over animated characters, novels and actors playing people who never existed.

I do not think the strongest future for synthetic performance is pretending that audiences cannot see the technology. Deception may create a short-term reaction, but it weakens the deeper experiment. If viewers know a character is synthetic and still return, still debate her decisions, still recognise her emotional patterns and still care what happens next, that tells us something much more interesting about storytelling. It suggests that emotional reality and biological reality are not identical.