It took me a week to get my head around everything presented at Google I/O 2019. So much has changed since I last went to the Shoreline Amphitheatre at Mountain View two years ago. It feels like Google has grown up.
Back then there was a lot of fun stuff— voice activated interfaces to order a sandwich or make a drink (literally — there was a machine with Google Assistant making a mixture for you), or to manage the A/C in your Audi Q8. The first consumer-ready AR/VR solutions had arrived thanks to the Daydream platform. Most of it was was fun, frivolous stuff for privileged people.
In 2017 Google announced that it was now an AI-first company. They debuted:
- Android Go — a system that works well on slower devices with limited data connectivity.
- Google.ai — a new home for artificial intelligence initiatives.
- TensorFlow Lite — new technology to run machine learning solutions on-device, without access to the Internet.
They were all very promising, yet very enigmatic.
I/O 2019 brought more clarity about Google’s AI-first mission: deliver tools to increase knowledge, success, health and happiness, and make them “available or everyone, no matter who you are, where you live, or what you’re hoping to accomplish”.
The solutions and technology that Google presented this year were focused more on solving real problems than impressing a few people in the Silicon Valley bubble.
Culture of inclusion
There are 500 million people in the world with speech disorders and hearing impairments that make verbal communication challenging. AI is making technology more and more accessible to people with such disabilities.
On-device AI can now create a live, accurate transcription of everything that it hears, including videos and podcasts. Thanks to Live Relay, people can make phone calls without using their voice or hearing, because their smartphone will speak and listen for them.
Project Euphonia is advancing voice recognition for diverse speech patterns and will improve the lives of people with impairments caused by neurological conditions, such as a stroke or traumatic brain injury.
Mobile technology is also becoming more inclusive. There are now 2.5 billion Android users in the world. Soon a further billion will come online. It is estimated that 80% of those people will come from developing nations across Asia and Africa, including Nigeria, India and Pakistan.
These users have very different digital needs and habits to people in the West. They are:
- Truly mobile-first. Most of them have never owned a desktop computer. Their digital revolution is happening entirely in the smartphone era.
- Likely to need more offline services because of unreliable WiFi and limited mobile data plans.
- Multi-lingual. India recognises 23 official languages. Hundreds of languages are spoken in Nigeria.
There are several solutions and technologies that will help us Android engineers to address the needs of the next billion Android users:
- Android Go, an operating system for affordable entry-level smartphones
- Firebase Performance Monitoring, which helps us to measure the speed of your web and app.
- MLKit brings on-device, machine-learning powered translations. Translate between 59 languages with no internet connection.
- Google Play Store makes it easier than ever to decrease the size of your app by splitting it into modules which can be delivered on demand.
Is your mobile product ready to be adopted by next billion users?
AI for benefit of all
AI technology is no longer confined to closed labs and universities, it’s serving people all over the world.
The Google.ai initiative, Google AI Impact Challenge, gave grants to 20 organisations that are using AI to change the world. One project uses intelligent analysis of crops to help farmers reduce pesticide use and improve yield. Another uses a deep learning-based bioacoustic monitoring system to track rainforest health and detect threats.
AI can now forecast flooding and predict lung cancer in a patient, long before a doctor could diagnose the disease. Solutions like Federate Learning will protect people’s data by running distributed machine learning operations on users’ devices instead of sending raw data to data centres.
And Deepmind, the AI that defeated human champions of chess and Go, is now being used to help humanity rather than compete with it. Examples include increasing the predictability and value of wind energy, or significantly decreasing power usage at huge data centres.
Finally, AI techniques and hardware are becoming much more accessible, so you don’t have to be a Google-size company to solve important problems. New TPU machines, soon to be available in Google Cloud for just a few hundred dollars per hour, will bring supercomputer power to small businesses.
And if you would like to apply AI to your product, there is a guidebook for designing human-centered AI products: People + AI Guidebook.
Are you ready for the future?
In this post, I’ve merely scratched the surface of what was presented during Google I/O 2019. But the things I’ve described weren’t selected randomly. They all contribute to making the world a better place. They all build bridges, break walls and remind us that diversity isn’t a problem but an asset.
They are all in line with Azimo’s mission — to democratise financial services, making them affordable and available to all. To improve millions of people’s lives. This is the future we’re preparing for. This is the future we want to build.