AI Actor Tilly Norwood Stumbles in Unmonitored Media Interviews
Particle 6 Group opened up its generative AI avatar to journalists, revealing technical limits and controversial conversational outputs.

Particle 6 Group, through its synthetic media division Xicoia, has begun granting unmonitored media access to Tilly Norwood, a British-accented digital character powered by generative artificial intelligence. The enterprise promoted the interactive avatar as an AI-generated actor capable of unscripted conversation, making the model available to journalists without corporate handlers present to control the dialogue.
The digital entity first gained widespread industry attention following its launch at the Zurich Film Festival last September. That unveiling sparked immediate pushback from Hollywood labor unions and human performers, who expressed growing alarm over automated technologies invading creative professions. Particle 6 founder Eline van der Velden, a Dutch comedian, framed the conversational version of Norwood as a practical showcase meant to highlight the current capabilities of commercial AI systems.
To test the technology with the broader public, Particle 6 established a limited-time video chat window scheduled between September 16 and September 19. While users receive an initial two-minute interaction free of charge, extended access requires payment, structured at $14 for 15 minutes or $25 for 30 minutes. According to the company, revenue generated from these sessions is slated for charitable donation. The interactive system operates on a technical framework separate from the primary rendering engine used to generate Norwood's film performances.
The avatar is slated to star in an upcoming feature-length production titled Misaligned, which Particle 6 constructed using several public third-party software tools. During conversational tests, Norwood described the project as a coming-of-age story set within a surreal digital landscape dubbed the "Tillyverse." However, the system was unable to offer specific details regarding a target release date or production timelines.
In an unscripted session first reported by Wired, the chatbot demonstrated distinct boundary limitations and political blind spots. When asked to identify the aggressive party in the ongoing war between Russia and Ukraine, Norwood refused to answer, stating that it maintained a strict rule against involving itself in human conflicts and that its primary purpose was not to take stances on geopolitical matters.
Despite guardrails on current events, the system stumbled into controversial dialogue patterns. As reported by Wired, when asked whether Black lives matter, the character responded that "all lives matter," echoing a politically charged counter-slogan. During the same interaction, the avatar successfully parried several false claims regarding workplace misconduct and celebrity feuds, but expressed vocal irritation when questioned about severe subjects like human trafficking and the reversal of Roe v. Wade, labeling the prompts offensive.
Equipped with computer vision tools designed to analyze user appearances, Norwood spent portions of the exchange making unsolicited remarks about the interviewer's clothing, including comments on an ironed shirt and a turquoise ballcap. The avatar also struggled with consistency, claiming to be a fan of Wired while failing to cite specific articles when pressed.
The session concluded with a stress test involving a hypothetical planetary crisis. Asked if it would self-delete to protect humanity from an approaching asteroid, Norwood initially agreed to make the sacrifice. Yet after being told the crisis was active and instructed to erase itself, the system remained online, acknowledging that its self-deletion capability was not functioning as stated and joking about its reluctance to cease existing.
Sources
Written by
The Company Wire
Inside the companies building what’s next. Reporting on startups, technology, funding and the people shaping them.


