Hey, that’s certainly odd - do you know if the voice input was selected accidentally, and it captured background noise? I don’t know what voice model is under the hood, but sometimes, like text generation models, voice models can hallucinate too, resulting in odd output.