Not sure I'm following, are you saying that Google's backend won't be able to handle the increased search volume if android users started using voice actions as much as Siri? Perhaps, and Siri's somewhat shaky launch don't bode well in that regards. But that should be a simple problem to fix (throw more hardware at it).
Yes, I'm saying that Google's backend has a very difficult growth problem to match Siri. It will need:
- More intelligent system that will require more processing of requests.
- It will need to be more throughout and complete in the services in offers
- It will need to be much faster than Google' voice control is currently (I haven't personally used it but I've heard that it's not fun to use because of the delay).
Once all those are dealt with Google will get a lot more users using it a lot more, which means even more HW to keep it comparable. Now factor in how many more Android-based devices there are in use compared to iPhone 4Ses and you see a dilemma that is much easier for Apple to deal with, hence my suggestion of how Google could introduce a similar service.
Not with the same ease or with seeing a direct return on the investment.
Of course it responds to keywords. That's how language works. Siri is more versatile and can parse your natural language better than any other competing service. Language and intelligence between human's isn't perfect either, nor ever will be.
Human's get tripped up with language and ideas, too. I'm not sure why that would suggest it's any less useful or difficult for Google to implement when its worlds above what anyone else has and what iOS 4.0 had with local Voice Control.
edit: Does Google's data mining from their GOOG-411 service help them with this or will they need to buy a company the way Apple did? (still surprised Apple kept the Siri name). There are a couple options that are already on the Android Market (Speaktoit and Vlingo), but none seem to be very good right now, though how good was Siri in 2009 before Apple bought it compared to today?