There seems to be hesitation to acknowledge that iOS might include a local speaker recognition system, even though it’s used in HomePod.
I understand that reading articles takes time so that is why I am posting excerpts.
I would just like to point out that I posted 2 different articles:
The one from 2023 references "the voice trigger system" in its conclusion, indicating that with iOS 17, the voice trigger system will support two trigger keywords on most Apple device platforms. This seems to indicate that this is not just a research, but an implementation, for a specific iOS version.
That said, I agree there’s no explicit mention of speaker recognition in Apple’s marketing or support materials, and iOS settings seem intentionally vague to avoid setting user expectations.
Regarding the HomePod vs. iPhone argument, the difference in threat models is noteworthy. The HomePod, as a stationary home device, may justify speaker recognition for sensitive information, while the iPhone, being mobile, might prioritize broader accessibility. Of course, this is speculative.
Finally, when contacting Apple Support, the guidance is not that “Siri will answer to anyone’s voice” but rather to retrain the local model, which suggests some degree of personalization is at play.
I’ve shared all my findings and sources to provide clarity, not to incite debate. I’d welcome more official clarification from Apple beyond their research articles, but without further input, I consider this discussion complete.