AI-avatar Tilly Norwood suffered a critical failure during an interview

Digital avatar Tilly Norwood, designed for public appearances and press interactions, encountered a technical glitch during a recent press tour. During one of the interviews, the model suddenly stopped responding to questions in English and switched to Chinese, raising questions about the system's stability.

The incident occurred live when the natural language processing algorithm demonstrated uncontrollable behavior. According to TechCrunch, the developers have not yet provided a detailed report on the causes of the "hallucination" or the failure in the language package switching logic. It is currently unclear whether this was an error in the model's weights or a failure in the prompt engineering responsible for contextual adaptation.

Such incidents highlight the risks of using autonomous AI agents as official company representatives. A sudden change in language or loss of speech coherence in the public sphere causes reputational damage and raises questions about the readiness of current architectures for full-fledged human replacement in media communications.

The technical community is awaiting the publication of system logs to understand whether this is a one-time bug or a systemic vulnerability in the avatar's architecture. Until the circumstances are clarified, the use of similar solutions in critical communications may be limited.