More and more, smartphones and other devices are using the voice-friendly tech to make life easier for users. Similar to Plum’s IVR, the software used by companies like Apple and Samsung draws off of a database of human voice recordings to determine what the speaker is saying, convert that into text and respond appropriately.
It seems like an easy enough process, but there are limitations.
The biggest challenge confronting speech-recognition software is the variance between individual speakers based on accent and other factors. In order to fit its software better for each individual user, iPhone’s Siri speech tool now personalizes its speech recognition by storing unique speech samples.
It works like this:
You pick up your iPhone and ask Siri, “What’s the weather like today in Boston?” Apple records your question and stores your vocal sample in their cloud.
The computers then analyze the speech and convert it to text; Siri interprets the question, searches the weather forecast for New York and responds back with a full report. The same way an IVR system gets info from a database for a caller.
That stored vocal sample stays with Apple for an indefinite period of time and helps them to better understand your vocal commands. The next time you ask what the weather is like in Boston, Siri will be able to understand you more quickly and recognize similar questions with ease.
Effectively, Siri learns your unique voice. Not only is the smartphone itself becoming more personalized, but so is Siri.
Stay tuned for the rest in “Privacy Matters”…