AI on your lock screen

For the last 10 years, news feeds have been the main way — the mainstream user interface — to discover interesting and relevant digital content. Today, news feeds, from Facebook and Twitter to LinkedIn, Instagram and Pinterest, are surfacing the interesting news and moments from your social network and favorite sources.

This is about to change. The push notifications on the lock screen of your personal mobile device are turning your lock screen into the new news feed. The lock screen is thus becoming the pivotal interface to access and experience any of the updates and content that you consider to be worth noticing.

Therefore, your lock screen and your mobile device, not the apps, become the nexus for all the personal data flows, feeding machine learning algorithms soon running also on your personal hardware.

This is a fundamental change. It will change the way your digital experience is personalized. It will change the way AI systems can learn from you. And it will change the power balance between the big industry behemoths such as Facebook, Google and Apple.

Battle for the lock screen

We’ve had push notifications bubbling under for some time now. Back in 2014, Christopher Mims of The Wall Street Journal predicted a big success to the Yo app because of the way it used the simple power of push notifications.

Yo didn’t rise to the occasion, but the applications and influence of push notifications has been growing ever since. Today, the landscape for push notifications is changing rapidly. Both Android and iOS have introduced updates on push features in a considerably fast pace.

Notifications are transforming from simple text-based boxes into adaptive elements that allow a richer and more nuanced experience, thus-called “rich notifications.” Designers and developers are embracing these new possibilities, enabling a more engaging user experience. Today’s notifications can contain text formatting, bigger images, video and updating infographics, as well as interactive features such as sharing. As a result, users are consuming more and more content directly on their lock screen.

The lock screen has become the place where your attention needs to be caught. And thus, every app is racing to invent more meaningful and engaging notifications. Nic Newman from Oxford’s Reuters Institute calls this “the battle for the lockscreen.” In the process, applications are turning into micro-platforms that can provide notifications as branded and optimized mini-products.

From rich notifications to smart notifications

The new richer interactions on your lock screen presents a new user interface paradigm and will have a major affect on personalization.

By appearing automatically on your lock screen, push notifications enable interesting things to find you, rather than the other way around. At the same time, the lock screen isn’t tied into presenting things in a chronological order. Push notifications allow you to experience things ambiently: notifications materialize on your lock screen automatically without your explicit action.

Importantly, you do not need to open the app to access content. Today, notifications from a news app allow you to follow the developing news event directly on your lock screen. You can participate in a conversation, check photos, watch a live video and share content without opening the app.

With a personalized lock screen feed your device has the potential to get truly smart and personal.

As we’ve seen in the news feeds of Facebook, Twitter and Pinterest, personalization algorithms are needed to curate the continuously growing flow of updates. Soon your lock screen is filtered by personalization algorithms, too.

Already, as people’s interactions are moving from the apps to the lock screen, both iOS and Android have started to automatize the way things are presented and accessed on the lock screen. Android provides automatically triggered smart notification bundles that collect together useful notifications. iPhone highlights apps based on your personal context, such as time and location. On both platforms, widgets are part of this development, serving richer interactions and more content without opening the app.

As an extension of you, your personal mobile device contains all your apps, thus making it a treasure trove of personal data. As an interface, the lock screen makes it possible to combine the data of your app-specific interactions with the rich contextual data provided by your device.

Concretely, the lock screen will introduce a new algorithmic layer for personalization. The lock screen captures your social interactions and content consumption patterns, your favorite apps, movies, videos, music and much more. This rich data will be used to feed machine learning systems to make personal suggestions and recommendations more relevant and contextual.

Soon your lock screen will filter push notifications actively and automatically, deciding which updates, suggestions, messages, apps, movies, recipes — and ads — are visible to you. With a personalized lock screen feed your device has the potential to get truly smart and personal.

From smart notifications to personal AI

When the interactions on your lock screen become richer, the data they generate becomes richer, too. Your mobile device will learn from everything you do more accurately than ever before.

This introduces a new opportunity to start really understanding you as a unique individual and thus go beyond the existing personalization gaps. Any individual app, even Facebook, couldn’t and can’t achieve this today (note: Facebook tried unsuccessfully to create their own mobile device; so did Amazon).

Personal hardware is becoming an essential part of personalization and machine learning. Today your mobile device uses various sensors and extensions — from camera to motion detector and smart wearables — to collect even more detailed data on you and your environment.

As Gary Marcus, the founder of Geometric Intelligence and NYU professor has pointed out, AI systems should be able to learn from a lesser amount of data. They should be able to learn like a child, continuously, iteratively and from everything, being able to generalize, apply and extrapolate these learnings in a useful way.

The personalized lock screen creates a unique interface connecting you and your personal AI running on your most personal device.

What if the missing piece for creating such a machine learning system has been a personal AI — an algorithmic angel, if you will — living and running on your most personal hardware, thus being able to learn with you like a child would.

Such a personal AI running and evolving with you on your personal device is taught and fed continuously by your rich interactions and contextual data. It evolves by iterating itself based on your feedback and personal patterns.

While learning directly from you, personal AI can utilize specialized internal and external agents that inhabit various digital environments, simultaneously utilizing the computing power in the cloud. In addition, these individual agents can process and provide domain-specific data, information and recommendations, from stock market tips to optimized travel options. The best versions of your personal AI collaborate and compete to evolve into better versions of themselves.

Who controls your personal AI?

Everything that happens on the lock screen is captured and can be used to enhance your experience — not directly by Facebook and other apps, but mainly by Google and Apple. Google Assistant and Siri will get smarter faster.

Google is already bringing machine learning into their devices using their own algorithms and hardware. Simultaneously, they are offering developer tools to optimize notifications. Apple is following suit. Samsung is trying to keep up with its recent acquisition of Viv, “the next-gen AI assistant.”

Will personal AI become your algorithmic angel, making sure you maintain your personal agency in tomorrow’s algorithmic reality? Or will it just turn your personal device into an ultimate marketing experience, thus trying to affect every decision you make?

The new age of personalized lock screen and personal AI makes the idea of algorithmic angels, your personally controlled algorithms, more timely than ever. Ethical committees and clauses are starting, but they don’t suffice. As our decision-making is augmented by intelligent machine learning systems, we need explainable algorithms, interfaces and methods to guide and control these smart entities in an explicit and comprehensible way.

The lock screen as a user interface provides a new interface to do so.

What if you swipe far enough left on your iPhone to see the settings and preferences of your personal AI? What if you can access various versions of these AIs and decide which one is active for a particular moment just by swiping your lock screen? Maybe you can have a mundane chat about the reasoning behind your AIs suggestions, or then you use intuitive gestures, haptics and sounds to communicate with each other in a mutually comprehensible manner.

The ultimate “conversational UI” won’t be an app or a bot that you need to open or call for. It’s something that’s present and available all the time, engaging in a continuous “dialog” with you and your digital and physical environments.

The personalized lock screen creates a unique interface connecting you and your personal AI running on your most personal device. This opens up completely new opportunities for designing next-generation human-machine communication methods and interfaces that can be applied from mobile devices to AR and VR environments. Simultaneously it is the next step to augment human and machine thinking in an inseparable way.