Can Apple’s SIRI work well with international user’s accent?
filed in Ecosystem iOS, Productivity Management on Oct.07, 2011

The Apple iPhone 4S presentation was largely a disappointment for a lot of folks, but for me the highlight was SIRI, the personal digital voice assistant.
Voice recognition is not something new. Google’s Android have this in their Android to replace text input or carry out search.
But Apple’s presentation looks really neat. Apparently I found out that SIRI exist as a startup before Apple decided to integrate this into their OS after buying over.
But can Android do something similar as well?
- It integrates narration of incoming text messages when put into a narration mode. Android has that too but its not integrated this way.
- Replies to text message through voice. Android is not integrated this way yet.
- Switches application context from text messaging to music playlist through voice command.
- Understands that you are currently in an application context and continues to be in that application context.
Till today Android have VLingo which does the same thing. It is free over at Android Market and well rated.[Get it here]
SIRI actually marries 2 concepts
- Instead of text commands to invoke switching between apps, it users voice. For text command apps on the desktop do take a look at Google Calendar’s Quick Add Feature [illustration here]
- SIRI maintains a set of voice commands. These voice commands could be 5 different voice sets to bring you to a single application context such as Reminders or Calendar, as we bring it up differently.
- SIRI then marries this with your own variable data input.
- The second powerful concept is the AI to recognize voice and translate to data or commands better than the competitors
Why SIRI is better than Android’s current offering is the AI portion is really well worked out.
- SIRI started off in 2003 as CALO, a DARPA funded project with 300 researchers. It has 8 years of development and a lot of money invested. Basically Apple is now bring a really advance voice to text and text to voice AI system to the mainstream consumers.
- The system will clarify after voice input to make sure you send or input something according to what you want.
- The conversion from units or time zones is going to be a very big feature.
- Integration with Wolfram Alpha as well.
The question now is that as Asians, not all of us have English names. In past voice recognition systems this have been a problem.
Will SIRI be a racist application? We will only find out when we get hold of one and try it.
The next question is: How is Microsoft and Google’s technology in this voice recognition area? Will they be able to compete?





