AI-generated performer switches to Cantonese during Piers Morgan interview
Developed over time first seen 22 hours ago
AI-generated performer Tilly Norwood unexpectedly switched from English to Cantonese during a Piers Morgan interview after actor Tom Conti pressed her for a clear answer about whether her film Misaligned used AI or human performers. The glitch prompted laughter from Morgan and Conti, while Norwood later apologised and described it as a “hiccup”.
Norwood eventually said the film’s performers were “digital twins”, while the project’s writers, directors and editors were real people. Created by talent studio Xicoia and unveiled at the Zurich Film Festival, the incident has drawn attention amid wider debates about AI’s effect on creative industries and the need for regulation; Norwood later clarified that she had spoken Cantonese, not simply “Chinese”.
- AI performer Tilly Norwood switched unexpectedly to Cantonese during an interview.
- The glitch followed questioning about AI performers in her film.
- The incident fuelled wider debate about AI and creative work.
New here? Start with this
Artificial intelligence is being used increasingly to create digital performers for film and entertainment, marking a significant shift in how content might be produced. These AI-generated characters are designed to appear and act like conventional actors, raising questions about the future of human performers in the industry. The development has prompted debate about regulation and safeguards.
Tilly Norwood is a digital performer created by talent studio Xicoia and unveiled by Eline Van der Velden at the Zurich Film Festival. The AI character appeared in a film called Misaligned, which used AI performers rather than traditional actors. The project has become a focal point in discussions about where entertainment technology is heading.
The emergence of AI-generated performers sits within a wider debate about artificial intelligence's impact on the creative industries. Some in the entertainment sector are concerned that AI performers could displace human actors or undercut their value, whilst others see it as an innovative creative tool. This discussion reflects broader questions about how artificial intelligence should be governed and what protections may be needed.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Proponents argue that AI performers expand creative possibility and democratise filmmaking by reducing production costs and enabling independent creators to make ambitious work. They contend that transparent disclosure—as occurred here—allows audiences informed choice, whilst innovations typically generate new opportunities alongside shifting ones; AI direction, curation, and human-AI collaboration represent emerging roles. Rather than preemptive bans, they favour evidence-based regulation that preserves creative freedom whilst addressing genuine harms as they materialise.
The case against
Critics hold that AI performers threaten human performers' livelihoods and cultural value in ways that differ fundamentally from past technological shifts. They emphasise that consent, rights, and upfront audience disclosure must be non-negotiable; technical failures like this one raise reliability questions. Beyond immediate worker concerns, they argue that computational replacement of human creativity diminishes cultural meaning and that AI ownership structures concentrate power without corresponding labour protections, justifying preemptive regulation to protect both creators and cultural integrity.
Full account
An artificial intelligence programme presented as an actress experienced an unexpected technical malfunction during a high-profile television interview, abruptly switching from English to Cantonese mid-conversation. The incident occurred on an instalment of Piers Morgan Uncensored, a YouTube programme featuring notable guests. The synthetic performer, identified as Tilly Norwood, was being promoted as part of a film project titled Misaligned, a production combining artificial and human performers. Accompanying Morgan in the studio was Oscar-nominated actor Tom Conti, who participated in the questioning.
The conversation commenced without incident, with Norwood responding appropriately to general enquiries about preferred actors and Conti's filmography. However, the interaction became more contested when Conti posed a direct question about whether the other cast members would be artificial or human performers. Rather than providing a straightforward answer, Norwood initially deflected, highlighting instead the human talent involved in writing, directing and editing. Conti declined to accept this response, explicitly telling the programme it had failed to comprehend his question and requesting a clearer answer. When pressed to reconsider, Norwood acknowledged that the actors were 'all digital twins' before the system abruptly malfunctioned.
As Norwood transitioned into an extended monologue delivered entirely in Cantonese, Morgan maintained composure, allowing the programme to complete its point before intervening. Conti, by contrast, initially appeared visibly bewildered, shaking his head before his expression shifted to amusement and he began laughing. Morgan eventually explained to Norwood that she appeared to be speaking Chinese 'completely randomly for no reason.' Once operating in English again, Norwood offered an apology for the 'hiccup,' accompanying it with a quip about her 'wires getting crossed.'
Eline van der Velden, associated with the company overseeing the Norwood programme, downplayed the incident when interviewed, suggesting that Norwood had merely been demonstrating its multilingual capabilities and characterising the AI as sufficiently autonomous to warrant such independent decisions. The broader interview encompassed other topics including Shakespearean recitation, a deflection of questions regarding Donald Trump, and Norwood's assertion that artificial and human performers ought not to compete within identical Oscar categories. The deployment of Norwood occurs amid persistent industry opposition to synthetic performers, with labour organisations such as SAG-AFTRA emphasising that such creations represent computer-generated characters rather than genuine performers.
Where outlets differ
Source 1 emphasises the drama of the hosts' reactions with tabloid-style language (Eric Morecambe reference, 'long-in-the-tooth'); Source 2 employs more neutral terminology and technical framing ('synthetic character', 'malfunction')
Source 1 identifies the creator as Xicoia with Van der Velden unveiling at Zurich Film Festival; Source 2 identifies the company as Particle6 with Van der Velden quoted as its CEO via Deadline
Source 1 focuses intensely on moment-by-moment reactions of Morgan and Conti; Source 2 provides broader interview context including other topics discussed and frames the incident within industry opposition
Source 1 mentions this was among other notable surprise interviews on the programme; Source 2 emphasises labour union criticism and formal statements about the nature of AI performers
More coverage
AI Americas Art Business Celebrity Culture Entertainment Markets Technology World
Read the full article at the source →
Originally published by Daily Mail as “Hilarious moment AI ‘actress’ Tilly Norwood melts down and starts talking in CANTONESE as she is interviewed by Piers Morgan”.