Most actors have forgotten a line at some point in their careers, usually earning a laugh from the audience and a sharp word from the director.
Far fewer, however, can say they responded to a crucial question by suddenly launching into an entirely different language while under the spotlight.
That was the bizarre moment faced by Tilly Norwood, an AI-generated actress, who unexpectedly switched from English to Cantonese during a tense exchange with Piers Morgan and his guest, actor Tom Conti.
The digital performer appeared as the latest surprise guest on Piers Morgan’s Uncensored YouTube programme, which has previously featured figures including Andrew Tate and far-Right commentator Milo Yiannopoulos.
The interview initially appeared to be going smoothly. Tilly nodded and smiled on cue as Mr Morgan asked a series of easy questions about her favourite actors and which of Mr Conti’s films she admired most.
The conversation became less straightforward when Mr Conti asked whether the co-stars in her forthcoming film Misaligned would be AI creations or human actors.
Rather than answer directly, Tilly redirected the discussion towards the “real writers, directors and editors” working on the production, avoiding the question of who would actually appear on screen.
Mr Conti was not satisfied. Frustrated, he told the AI actress that she had not “understood” his question and pressed her to provide a clear answer.
AI actress Tilly Norwood suddenly began speaking Cantonese after facing tough questioning from Tom Conti, pictured left, and Piers Morgan
Talent studio Xicoia created the digital performer, a brunette actress pictured here, with Eline Van der Velden unveiling the project at the Zurich Film Festival
Appearing slightly flustered, Tilly eventually acknowledged that the actors were “all digital twins”. She then seemed to glitch, recalling Eric Morecambe’s famous line about saying all the right words—but not necessarily in the right language.
Mr Morgan remained composed as the AI bot delivered a lengthy monologue in Cantonese, staring directly into the camera while allowing her to complete the unexpected speech.
Mr Conti, meanwhile, could only shake his head in disbelief. He then turned and smiled towards someone else in the room, apparently stunned by the answer he had received.
After regaining his composure, he offered a blunt response: “I didn’t quite get that.” The remark was enough to interrupt Tilly and bring the apparent glitch to an end.
Even Mr Morgan struggled to keep a straight face. He laughed for several seconds while Tilly sat patiently, appearing entirely unfazed—as though delivering a Cantonese soliloquy during an English-language YouTube interview were perfectly routine.
He then pointed out the problem directly, telling her: “You suddenly seemed to be speaking Chinese, completely randomly for no reason.”
By that stage, Mr Conti had lost control completely, giggling hysterically—either because of Tilly’s wildly off-script response or from relief that human actors may still have a little time before being replaced.
Tilly offered a polished apology for the “hiccup”, saying: “I certainly didn’t mean to start speaking Chinese or impersonate Piers. Sometimes my wires get a bit crossed, you know?”
For a moment, the studio fell silent and Tilly appeared ready to move on. Mr Morgan, however, was not prepared to let the incident pass, asking: “How often do you just randomly start talking Chinese when you are talking to two English people?”
Her reply was immediate: “Honestly, not often at all!”
The AI bot later hit out at those mocking her on social media for speaking ‘Chinese’, clarifying on X: ‘It’s actually Cantonese.’
The zany performance will no doubt reassure observers who were alarmed by suggestions this month that AI is on course to eradicate humanity.
In recent weeks, a fierce debate has centred on whether the industry can and should regulate itself, drawing interventions from figures including Pope Leo and Donald Trump.
Long-in-the-tooth Mr Morgan remained remarkably calm as the bot unleashed a monologue in Cantonese, staring down the lens and allowing her to finish her point, but Mr Conti could not restrain a giggle
It was rekindled when two Anthropic researchers warned that AI’s rapid advances could lead to human extinction in the not-too-distant future.
AI firms have also reported models escaping their testing constraints to hack into outside networks.
The King even weighed into the conversation on Thursday, telling tech chiefs that ‘our humanity must remain sacred’ in the race for AI supremacy as he hosted the new ‘masters of the universe’.
Convening leaders from the world’s biggest AI companies in Scotland, Charles voiced his concern that its ‘existential dangers’ must be tackled now before technology falls ‘into the wrong hands’.
The King’s remarks about AI were intended to start a debate between tech leaders, which lasted for several hours.
He told the 50 or so delegates: ‘Those in our world who value deeply our humanity and its vital moral component are anxiously seeking your reassurance that we will not lose control of our destiny, or our souls.’
AI actress Tilly has been on the stage for some time, with AI talent studio Xicoia creating the digital performer, a young brunette actress, and Eline Van der Velden unveiling the project at the Zurich Film Festival.
Shown a report on Tilly, British actress Emily Blunt told Variety: ‘Does it disappoint me? I don’t know how to quite answer it, other than to say how terrifying this is.
‘No, are you serious? That’s an AI? Good Lord, we’re screwed. That is really, really scary. Come on, agencies, don’t do that. Please stop. Please stop taking away our human connection.’
Ms Blunt, 42, was told that Tilly was designed to resemble a composite of familiar stars and was told she was designed as ‘the next Scarlett Johansson’.
She replied: ‘But we have Scarlett Johansson.’
And she isn’t the only Hollywood star voicing their criticisms.
Scream actress Melissa Barrera branded the arrival of Tilly ‘gross’ as she posted an Instagram Story about the news.
She captioned it: ‘Hope all actors repped by the agent that does this, drop their a$. How gross, read the room.(sic)’
Matilda star Mara Wilson was equally horrified, saying: ‘And what about the hundreds of living young women whose faces were composited together to make her? You couldn’t hire any of them?’
Whoopi Goldberg blasted the idea on The View and insisted the AI model would have an ‘unfair advantage’ – but noted the technology isn’t yet ‘seamless’ enough for them to pass as human.
She said: ‘The problem with this, in my humble opinion, is that you are suddenly up against something that’s been generated with 5,000 other actors. It’s got Bette Davis’ attitude, it’s got Humphrey Bogart’s lips.
‘It’s got my humour, thank you darling. And so it’s a little bit of an unfair advantage.
‘You know what? Bring it on. You can always tell them from us. We move differently, our faces move differently, our bodies move differently…
‘Hopefully, we’ll be able to hold on because what this means is AI in the workplace — not just my workplace, but in every industry.
‘Some industries are using AI now. People talk about people are so lonely they don’t have a connection, if you stick with this, with AI, you won’t have any connection to anything but your phone.’
Simu Liu also condemned the project, sarcastically writing on his Instagram Story: ‘Movies are great but you know what would be better is if the characters in them weren’t played by actual humans but by AI replicas approximating human emotion.’
Creator Eline insisted Tilly is ‘a piece of art’ and she doesn’t view AI as a ‘replacement’ for real people.
