AI Avatar Tilly Norwood Glitches to Cantonese During Piers Morgan Interview

Key points

• Norwood switched to Cantonese mid-sentence while answering a question about whether her co-stars in Misaligned were human or digital twins.
• The glitch was described by Norwood as a 'little hiccup' where her 'wires got crossed,' though Particle6 founder Eline van der Velden suggested it might have been an attempt to showcase language skills.
• The interview featured actor Tom Conti, who noted Norwood initially failed to answer his question about the nature of the cast before the language switch occurred.
• The viral clip has amplified concerns about the stability of AI models and the ongoing controversy surrounding the use of AI actors in the entertainment industry.

Background

Tilly Norwood, created by Particle6 in 2025, has been a polarizing figure in Hollywood since her unveiling at the Zurich Film Festival. Previous coverage noted that the AI avatar has faced condemnation from actors' unions and human performers, who view her as a threat to jobs and artistic integrity. Norwood has been conducting a publicity campaign for Misaligned, a hybrid production that mixes AI-generated characters with human directors and writers. Earlier reports indicated that the AI model has struggled with basic factual queries and political neutrality in other interviews, often deflecting questions or providing vague responses.

How outlets are covering it

HuffPost and Deadline focused on the technical failure and the viral nature of the Cantonese glitch, emphasizing the 'hiccup' explanation provided by the AI itself. Deadline included a comment from Particle6 founder Eline van der Velden, who framed the incident as a display of the AI's autonomy and language capabilities rather than a malfunction. In contrast, WIRED provided a broader critique of Norwood's functionality, detailing a separate interview where the AI failed to engage with serious topics, offered nonsensical answers, and exhibited erratic behavior. WIRED highlighted the absurdity of the AI's responses, such as its refusal to acknowledge geopolitical conflicts and its fixation on the interviewer's clothing, suggesting that the Morgan interview glitch is part of a larger pattern of unreliability in the AI's performance.

Why it matters

The incident underscores the risks of deploying AI models in high-stakes, public-facing roles without robust error-checking mechanisms. As AI technology advances, the potential for unpredictable outputs in live broadcasts or interviews poses challenges for media reliability and public trust. The viral nature of the glitch also highlights the public's skepticism and amusement regarding AI integration in entertainment, potentially influencing how studios and networks approach the use of digital talent in the future.

What to watch

Particle6 is expected to address the technical issues behind the Cantonese glitch, possibly through software updates or improved training data. The entertainment industry may see increased scrutiny of AI actors as unions and human performers continue to oppose their use. Future interviews or appearances by Norwood will likely be monitored for similar glitches, and the public's reaction to these incidents may shape the acceptance of AI in mainstream media.