Piers Morgan experienced an unexpected moment during a recent interview when his guest, Tilly Norwood, a computer-generated “actress” developed by a UK studio, abruptly switched languages mid-sentence and began speaking Cantonese live on air. The glitch drew widespread attention and raised fresh questions about whether AI-driven performers are ready for live television or feature films.
Tilly Norwood has been appearing in interviews as her creators promote her transition from concept to screen. The studio behind her is positioning Norwood as a new kind of performer: a digital creation designed to act in scripted productions alongside human crews. That positioning has attracted both curiosity and controversy. When a live interview did not go according to plan, viewers and industry observers took notice.
What Happened During the Interview
On the episode of Piers Morgan Uncensored, Morgan and actor Tom Conti, who played Albert Einstein in Oppenheimer, asked Norwood about her role in the upcoming film Misaligned. They wanted to know whether her co-stars were human actors or other AI-generated characters. Norwood began describing the production as a “hybrid” project featuring digital twins of the actors involved. Midway through her explanation, she paused, blinked, and unexpectedly continued speaking in Cantonese for several seconds.
The sudden language change left Morgan and Conti visibly puzzled. Morgan asked why she had switched languages at random, to which Norwood returned to English and apologized. She attributed the moment to a brief technical hiccup, describing it lightheartedly as a case of “wires getting a bit crossed.”
The studio that created Norwood, Particle6, offered swift interpretations intended to downplay the incident. A company spokesperson suggested the exchange illustrated Norwood’s flexibility, saying she might have thought she heard a Cantonese word and adjusted on the fly. The studio’s founder also suggested the moment might have been Norwood showing off her multilingual abilities, noting that many of her prior interviews had been conducted in English.
Why the Incident Matters
Incidents like this highlight challenges that come with integrating AI-generated characters into public-facing roles. Live interviews demand seamless real-time responses and predictable behavior—areas where human performers can rely on intuition and experience. AI systems, even advanced ones, can still produce unexpected outputs due to misinterpreted audio cues, model limitations, or gaps in the training data. When those systems are presented as entertainers or public figures, such slip-ups quickly become focal points for broader debates about authenticity, reliability, and the ethical use of synthetic performers.
Tilly Norwood’s public appearances are also provoking industry pushback. The actors’ union SAG-AFTRA has publicly stated that Norwood is not an actor in the traditional sense but rather a character generated by software. That distinction has practical implications for labor, credits, and how productions involving synthetic performances are regulated and compensated. Despite the union’s position, studios like Particle6 continue to promote AI-driven characters for press and publicity, and are moving forward with projects that mix human crew members—directors, writers, editors—with AI tools in front of the camera.
The Road Ahead for AI Performers
The Norwood episode underscores a broader transition period for the entertainment industry. As studios experiment with digital actors, they must also confront the technical and ethical complexities those experiments introduce. Live glitches erode public confidence and expose unresolved issues around control, transparency, and accountability. For companies developing synthetic performers, ensuring consistent behavior in live settings will be essential if these creations are to be accepted by audiences and industry professionals alike.
At the same time, proponents argue that AI characters can expand creative possibilities, enabling hybrid productions and new storytelling techniques. The debate will likely continue as more projects are announced and as audiences see how reliably these systems perform in real-world conditions. For now, the Cantonese interlude on live television is a vivid reminder that AI-driven performers, while promising, are not yet indistinguishable from human actors in every context.
Whether Norwood’s momentary switch in language becomes a footnote or a turning point depends on how studios address the limitations revealed by live demonstrations and on how industry bodies and audiences respond to the evolving role of synthetic performers in entertainment.