Smartphones with enhanced capabilities
- Canara Bank plans to hire bankers for up to Rs3,500 crore QIP
- Zohra: Inspiring sounds of music in war-torn Afghanistan
- Devendra Fadnavis govt may face fury over farm distress during winter session
- Security forces pin hopes on public ire against J&K militants
- RIC meet: Foreign ministers of Russia, India, China meet today to boost Asia-Pacific relations
Artificial Intelligence (AI) is undoubtedly the biggest trend in smartphones at present. It can respond to voice commands, text messages and be part of an existing conversation within an instant messenger app. It can offer contextual suggestions, respond to queries and execute requests, among other things. And it can do all this because it learns, discovers, specifies and generalizes along the way to learn different usage scenarios and individual preferences.
AI has been around since the 1990s, though with products such as speech-to-text software, which were more limited. Over the years, it has evolved, become more precise, reaching a point where it truly can assist humans. For instance, the Google Home assistant, according to Google, now has a 4.9% word error rate in speech recognition, down from the 8.5% error rate in July 2016.
One focus now is smartphones, which allow us to stay in touch with friends and family, access apps, handle emails, get navigation assistance, etc. Integrating AI into their different apps and functionalities is bound to improve that interaction.
It is no surprise then that smartphone makers like Huawei are putting in time and money to develop a full-fledged AI-powered smartphone by 2020. By then, the smartphone-installed base for AI will exceed six billion units (up from four billion in 2016), with revenue touching $355 billion (around Rs21 trillion), according to estimates from a white paper released by research firm IHS Markit in March.
If you thought Apple’s Siri, Microsoft’s Cortana and Amazon’s Alexa were amazing, you’ve seen nothing. We look at all that AI enables today, to try and understand what we can expect in the future.
A context to everything
The main challenge for AI tools has been enhancing their capabilities to understand context. Google’s Assistant, which started its journey within the Allo messenger app (free on Android and iOS), is now a full app on Android smartphones. It can understand the context well. For instance, ask “what are the best Italian cuisine restaurants in Delhi” while you are chatting with a friend, and it’ll show suggestions within the same chat window. Assistant can also set reminders, help with translation, answer questions, check the weather forecast and even control smart-home gadgets.
The Samsung Galaxy S8 and S8+ phones have introduced an AI tool called Bixby. It understands your routine and adds context—in the morning, you’ll get your calendar tasks; in the evening, it’ll be replaced by a personal to-do list and nearby places for coffee, for example. Bixby also plugs into third-party apps such as Twitter. The limitation is that Bixby doesn’t have voice-command capabilities in a few countries, including India (these should be available soon though).
Similarly, Sony’s Xperia XZ Premium smartphone runs the Xperia Intelligence Engine. It will understand how you use the phones, the apps you prefer, and will suggests tips and options based on app usage—how often you use an app, what time of day, and more.
Know you better
The critical ingredient of any AI tool is the ability to understand usage patterns and modify its behaviour.
HTC’s AI tool, Sense Companion, is available in the U Ultra and the forthcoming HTC 11 smartphone. This will remind you about forthcoming engagements in your calendar, the weather conditions before you travel, track your fitness, and send an update when the battery is running low. Sony’s Xperia Intelligence Engine sends customized suggestions, such as managing connectivity options while travelling.
The follow-up machine
When you have AI on your phone, it cannot sit in isolation. Panasonic has embedded into its custom interface for Android phones, a smart AI assistant called Arbo. Among its various features, Arbo includes the ability to pull location data from your phone’s GPS, so it will know once you are in office, and based on your calendar, can automatically set up a video call for you. It’ll also remind you to return calls you may have missed. Like most AI tools, it’ll learn with time—including which apps you prefer at which time of day or at which location. At present, the Eluga Ray X and Ray Max phones have Arbo.
The ecosystem chain
It is not just phone manufacturers who are bolting on the intelligence. Google is rolling out updates for the popular Photos app (free; Android and iOS), with enhanced AI that will identify the friends in those photos. The new Google Lens app will run a complex algorithm on a photo, and will be able to pull data from the image in the form of text—one example is the ability to point the phone’s camera at an event information display; the information will be added automatically to your calendar.
All the augmented reality smartness cannot run without powerful hardware to drive the algorithms. Qualcomm’s flagship Snapdragon 835 chip is optimized for machine-learning apps—it reduces the need to access the cloud and lets critical processes run on the processor and the graphics chip. Qualcomm explains, for example, how an audio or speech detection application might run on the Hexagon digital signal processor (DSP) hardware in the phone, while a graphics app running on the same phone can utilize the Adreno graphics hardware.
So there is a lot to look out for, on both the hardware and software fronts.