Siri, Google Now and Cortana: How digital assistants predict what you need

In October 2011, Apple announced its digital assistant Siri as a feature on the iPhone 4S. Soon it became popular for its witty replies, rather than the ease of accomplishing a task. The next year in July 2012, Google Now made a debut on Android smartphones. It was known for the ease with which it integrated with Google services and enabled tasks using cards that helped you understand what you need to do next. Gathering data Typically, digital assistants across the device platforms work by understanding as much about you as possible. This happens by accessing patterns and data from your email, track past GPS location patterns, as well as past requests made to it by the user. To top it all, digital assistants continuously learn from the millions of requests it receives from users across the globe. Machine learning, as engineers call it, enables this device to adapt its responses depending on the nature of queries. Recently on a popular TV show, Siri was asked by the host Stephen Colbert what he could ask his guest Tim Cook. The reply was a smart, “Do me a favor,” she responded. “Ask him when I’m going to get a raise.”

Humour and witty responses such as these require a great deal of intelligence that could come in from multiple sources. In this case Siri knew who Tim Cook was (we’d be surprised if she didn’t) and effectively threw a response that was very human. Humour, intelligence, or plain human What makes a digital assistant popular are such human responses with humour and intelligence. Earlier, Microsoft demoed how Cortana is learning Indian mannerisms. [caption id=“attachment_274799” align=“aligncenter” width=“416”] If Cortana was to share a joke for India, it would probably involve food! Image: Microsoft[/caption] Cortana, we are given to understand, would admire Sachin Tendulkar given his popularity in India. There was also mention of a joke that involved a naan. You could **read more on that here** . Active or passive assistance? In a detailed piece on digital assistants, search engine guru Danny Sullivan differentiates between active and passive assistance. Simply put, there are some tasks that are very repetitive in nature. For instance, you wake up in the morning, and depending on whether you are in a city such as Mumbai or Delhi, you either take the local train or head out on the road in the NCR region, waiting to crib about rush hour traffic. The situation isn’t any different across the other prominent cities of India. Hence, traffic updates help plan your time. That’s passive assistance. It does it without any input from you. However, let’s consider an instance when you have a sudden meeting out of office. That’s when an assistant would fetch information from your email or calendar and route you to the meeting destination and time and add information from traffic updates to give you a complete plan. Similarly, if you have a flight, and due to bad weather the flight is delayed, the information is factored in to create an itinerary for you. It simply isn’t humanly possible to invest time into sorting your plan for the day. Self help books would advice you to do all of it the day before, but we all know that we’re simply too lazy to get into such a discussion at the end of the day. Now that technology enables it effortlessly on a platter with notifications to ensure we don’t need to remember a thing, digital assistants have immensely risen in popularity. Understanding Digital Assistants In a report, University of Michigan, sheds some light on an open source project titled Sirius which is termed as an open source computing platform you control with your voice. The report describes digital assistants as robust applications that accept voice instructions and queries, process them to interpret them, and turn around a response in spoken words. In the following video, Jason Mars, Assistant Professor, Electrical Engineering and Computer Science College of Engineering at the University of Michigan explains what goes behind digital assistant.