Google I/O 2018 Keynote: AI Everywhere

Google‘s annual I/O developer conference is one of the biggest events on the tech calendar. It‘s where the company unveils its latest innovations to an audience of thousands of developers, press, and industry players.

While I/O features hundreds of deep-dive sessions over three days, the opening keynote is where Google‘s leadership lays out their high-level vision. It‘s a glimpse into what one of the world‘s most influential companies is prioritizing now and in the coming years.

This year‘s I/O keynote, delivered over two hours to 7,200 attendees and millions of livestream viewers, continued Google‘s transformation from a mobile-first to an AI-first company. From the Android operating system to Google Maps to Gmail, artificial intelligence is being infused into everything Google does.

Key I/O 2018 stats:

  • 7,200+ attendees on-site
  • Millions of online viewers
  • 3 days
  • 200+ sessions
  • 100+ product & platform announcements

It‘s All About Machine Learning

Google CEO Sundar Pichai kicked off the event by stating Google‘s overarching philosophy: "We‘re moving from a company that helps you find answers to a company that helps you get things done."

The path to accomplishing that is paved with AI and machine learning. Pichai shared that Google now considers itself an "AI-first" company and that they are "making a big bet on it" across the board.

This AI-first vision manifested itself in several ways:

  • A next-gen Google Assistant that can carry on natural phone conversations to handle tasks like booking appointments
  • New AI-powered features in Gmail like Smart Compose, which suggests complete phrases as you type
  • Adaptive Battery and Brightness in Android P that learns your usage patterns to optimize performance
  • Powerful computer vision capabilities in Google Lens and Google Maps
  • More intelligent news curation and recommendations in Google News

The through line in all these efforts is using machine learning to deliver more helpful, context-aware, and personalized experiences. As Pichai put it, "we are working hard to give users back time."

One of the most impressive demos was Google Duplex, a new system that enables the Assistant to conduct natural conversations over the phone. In a pre-recorded demo, the Assistant called a hair salon to book an appointment, carrying on a remarkably human-like dialogue complete with "umms" and "mmm-hmms."

While still experimental, Duplex points to a future where AI can handle ever more complex real-world tasks for us. That‘s equal parts exciting and unsettling, as it will surely raise new questions around AI ethics and transparency.

Digital Wellbeing & Meaningful Engagement

Another major theme of I/O 2018 was what Google termed "digital wellbeing" – helping people foster a healthier relationship with technology and their devices. Pichai acknowledged the "increasing concern about people spending too much time on their phones."

In response, Google is baking new features into Android P and YouTube to help people better understand and control their usage:

  • Android Dashboard: Provides a daily and weekly overview of how you spend time on your device, including time spent in apps, number of notifications, and how often you check your phone
  • App Timer: Lets you set time limits for specific apps, graying out the icon when the limit is reached
  • Wind Down Mode: Switches on Night Light and Do Not Disturb and fades the screen to grayscale at a selected bedtime to help you disconnect
  • YouTube custom break reminder: Set reminders to take a break while watching videos

Google also said they are "committed to meaningful engagement" with users, not just driving up time spent in apps and on devices. They pointed to Gmail‘s high-priority notifications and time-sensitive snoozing as examples of helping people manage their attention.

This focus on digital wellbeing reflects the techlash and concerns about tech addiction that have gained steam recently. By getting ahead of the issue, Google aims to position itself as part of the solution rather than the problem.

What‘s New in Android P

As usual, I/O brought a slew of announcements about the next version of the Android OS, code-named Android P. Developer previews have been available for months, but at I/O Google revealed more consumer-facing details.

Some of the highlights:

  • Adaptive Battery uses machine learning to prioritize system resources for the apps you use most
  • Adaptive Brightness learns how you like to set the brightness slider in different settings and does it for you
  • App Actions predicts what you want to do next based on your context and displays relevant action shortcuts
  • Slices shows interactive snippets of your apps in places like search results
  • New system navigation ditches the traditional home, back, and multitasking buttons for a single, gesture-driven "home" button
  • Digital wellbeing features like Dashboard, App Timer, and Wind Down mode

Under the hood, Android P has a variety of upgrades aimed at developers, like enhanced messaging, multi-camera API support, indoor positioning with Wi-Fi RTT, and more.

Android P Beta is available now on seven devices from Google, Essential, Nokia, Oppo, Sony, Vivo, and Xiaomi. The final consumer release will likely come in Q3.

Maps: The Real-World Interface

Google Maps is one of the company‘s most popular and mature products, but it‘s far from done evolving. At I/O, Google previewed a series of updates that position Maps as a hub for interacting with the real world.

A new For You tab provides a personalized, always-updating list of recommendations for restaurants, bars, and points of interest near you or in neighborhoods you frequent. It takes into account your preferences, location history, and context to surface useful suggestions.

Google is also using computer vision and Google Lens to make Maps more powerful. A VPS (visual positioning system) will help you orient yourself and navigate indoors. And soon you‘ll be able to point your camera at businesses and places to see helpful overlays of information like hours, reviews, and daily specials.

Even getting around is getting smarter. Maps will draw from its vast data set to provide predictive travel times – estimating how long driving, biking, or ride-sharing to a destination will take at different times of day. It will also predict parking difficulty and suggest alternative spots to leave your car.

Gmail Gets Smarter & More Secure

Google rolled out a major redesign of Gmail just a few weeks ago, and at I/O they doubled down with several new AI-driven features for the popular email service.

Smart Compose is like autocomplete for email. As you‘re typing, it will suggest full phrases that you can insert with a single tap. The suggestions are based on the context of your message and how you typically write.

Smart Reply, which offers three short responses you can send with a click, is getting an upgrade. It will now be able to parse attached photos and suggest replies referencing them (e.g. "The hike looks amazing!"). And on Android you‘ll be able to quickly view and tap Smart Replies from the notification shade without opening the app.

On the security front, Google is integrating anti-phishing protections into Gmail. If you receive a suspicious email, you‘ll see a big red warning banner explaining that it looks like a fake. Links to known malicious sites will also be flagged in red.

With over 1.4 billion active users, these changes to Gmail will reshape the email experience for a huge swath of people. The AI-powered features aim to increase productivity, while the security enhancements help fight the scourge of phishing.

Google Lens: A New Way to Search

Google Lens, first announced at I/O 2017, is a Shazam-like tool that lets you identify objects, landmarks, text, and more just by pointing your phone camera at them.

While it‘s been available for months, at I/O 2018 Google announced a major expansion of Lens. It will now be built natively into the camera apps on supported devices from LG, Motorola, Xiaomi, Sony Mobile, HMD/Nokia, Transsion, TCL, OnePlus, BQ, and Asus. That‘s in addition to being part of Google Photos and Assistant.

Google also previewed new capabilities for Lens, including:

  • Smart text selection: Copy and paste text from the real world into your phone
  • Style match: Point your camera at an outfit to see similar clothes, along with purchasing links
  • Real-time results: Lens will proactively surface information in the camera viewfinder

With the Lens button in their native camera app, hundreds of millions of Android users will soon have an intuitive new way to search the world around them. It‘s a taste of the physical/digital convergence that many tech futurists predict is coming.

Upgrading Google News with AI

In the wake of Facebook‘s algorithm tweaks deprioritizing news, Google is banking on its own AI chops to create a better news destination.

The revamped Google News app uses machine learning to understand the people, places, and things involved in a story and organize articles into storylines you can follow. It also factors in your interests and reading history to spotlight the most relevant content.

A new Newscast format presents each story as a collection of headlines, local news, FAQs, social commentary, and fact-checking info. The idea is to create a well-rounded view of a topic that evolves as the story develops.

On the curation front, the app‘s For You section pulls in a personalized briefing of five stories Google thinks you need to know that day. Full Coverage provides a complete picture of a story from a variety of trusted sources. And Newsstand makes it easy to find and subscribe to sources you trust.

With over 10 billion monthly page views for Google News, these AI-driven improvements could have a big impact on how people consume journalism. Of course, Google will have to be vigilant about surfacing authoritative sources and fighting filter bubbles.

AI for Social Good

Amid all the product announcements, Google made sure to highlight ways they are applying AI and machine learning to tackle societal issues.

For example, they are:

  • Developing an early warning system for floods that has already sent alerts to 100,000+ people
  • Using DeepMind to help diagnose eye diseases and predict heart attacks
  • Working with researchers to track and prevent the spread of infectious diseases
  • Partnering with the UN to monitor environmental changes with satellites

Google.org, the company‘s philanthropic arm, also announced a $25 million AI Impact Challenge to support organizations using AI to solve social and environmental problems.

It‘s heartening to see some of the tech industry‘s brightest minds and most powerful tools being directed at important issues beyond serving ads and keeping us hooked on screens.

More Self-Driving Car Progress

Last but not least, Google‘s sister company Waymo shared some updates on its self-driving car program.

CEO John Krafcik revealed that Waymo‘s autonomous vehicles have now logged over 9 million miles on public roads. He showed off the latest prototype with enhanced LiDAR, vision systems, and radar that can spot a person from 500 meters away.

Waymo is already operating a beta self-driving taxi service in Phoenix, Arizona. Krafcik announced that the program will expand to more participants in the coming months. And in a dig at Uber‘s recent high-profile accident, he emphasized that safety is Waymo‘s highest priority.

While self-driving cars are still years away from being mainstream, I/O 2018 showed that Waymo continues to lead the pack in making them a reality. With Alphabet‘s backing and Google‘s AI prowess, they are well-positioned to transform transportation in the coming decades.

The Takeaway

I/O 2018 made it abundantly clear that Google is going all-in on AI and machine learning. Nearly every announcement had an AI component, from adaptive battery management on Android to contextual recommendations in Google Maps.

Some of the previews, like Google Duplex carrying on a human-like phone conversation, were jaw-dropping. Others, like Smart Compose for Gmail, were less flashy but will likely be widely used. Regardless, the scope and ambition of Google‘s AI efforts is staggering.

At the same time, Google is clearly paying attention to the growing backlash against the unintended consequences of technology. The digital wellbeing features woven throughout Android P and other products are a proactive attempt to rebalance our relationship with screens and algorithms.

Google‘s embrace of AI makes sense given its unique strengths and market position. With billions of users across its platforms, the company has access to an unparalleled amount of data to feed its algorithms. Its vast knowledge graph understands connections between people, places, and things. And its deep bench of machine learning talent is constantly pushing the envelope of what‘s possible.

Still, the march towards an AI-first future raises major questions that I/O 2018 mostly glossed over. How will Google balance its insatiable appetite for data with growing privacy and security worries? What ethical lines will they draw in the sand as they give AI systems more autonomy and decision-making power? How can they apply the same algorithmic approaches to fighting issues like filter bubbles, toxicity, and misinformation?

The coming months and years will reveal how successfully Google can navigate these challenges. What‘s certain is that the concepts and products unveiled at I/O 2018 will touch billions of lives as they reshape how we interact with technology on a daily basis. Given Google‘s ubiquity and power, that‘s simultaneously thrilling and a little terrifying.

One thing‘s for sure: the AI-first era is here, and Google is determined to win it. Buckle up because the next few years will be a wild ride with major implications for the tech industry and society as a whole.