The problem with these devices is they don't ask for more context, like a human might. Sometimes they use telemetry for context, but most don't seem to keep any memory of previous questions asked.
If Siri asked: did you mean London, Canada or London, England, and then perhaps weighted that response in the future, that would probably end up creating a feedback loop where it would get quite good.
In the case of Siri, Apple has said they are very committed to privacy, and this might preclude this type of solution.
If Siri asked: did you mean London, Canada or London, England, and then perhaps weighted that response in the future, that would probably end up creating a feedback loop where it would get quite good.
In the case of Siri, Apple has said they are very committed to privacy, and this might preclude this type of solution.