AI and machine learning are about to be everywhere.
Google’s has been using AI (Artificial Intelligence) for several years, but Google I/O 2018 is about putting it in everything, everywhere. From your doctor’s office to the car you call to take you home, AI and Machine Learning are going to work behind the scenes and make the tasks you do every day feel like magic. Seriously.
What did Google announce about AI at I/O 2018?
- Healthcare — Google’s work in diabetic retinopathy has been expanded to detect cardiovascular crisis before it happens, help measure A1C levels, predict BMI, measure systolic blood pressure and more. Trials with physicians will begin this year.
- Gmail — Gmail will help you write by predicting what you need to write! AI looks at the subject and who you’re emailing and can help you finish your thoughts.
- Accessibility — Google demoed a CCTV system that can identify each speaker and isolate what they are saying, even in a heated conversation when everyone is talking over each other. And we saw how AI can work in Gboard to help people with disabilities that prevent them from communicating. It was wonderful to see.
- Photos — Suggested actions help you know the things you can do with the photo you’re looking at. Like sharing with the people in the photo or fixing the exposure or even “fixing” a document that was angled and unreadable in a photo. And Google Lens gets a monster of an update, too!
- Assistant — Assistant is getting more voices, and using AI lets them sound more natural and communicate more naturally. Seeing Google Assistant make a call and schedule a hair appointment by itself was amazing.
There are so many advancements in AI that Google was forced to add liquid cooling to the Tensor Flow data centers. You know things are serious when you need a radiator to keep a computer running.
How will that affect Android in the future?
The core of Android is getting its own changes with Android P, but AI is something that works better on individual tasks. The apps you already love and use can integrate Google’s AI technology to do more and (hopefully) do it better. We need to look no further than the Pixel and its camera here. A single, off-the-shelf lens camera was one of the best we have ever seen because Google uses AI to correctly adjust for the scene and in the post-processing. That’s not magic — it’s years of work happening all at once in the camera’s software in conjunction with the processor.
Android comes in to make everything work together. That’s not an easy task.
When will see any of this?
Soon. And later. And maybe never.
Some, like Assistant’s new voices that sound like real people, are coming right away. Photos and Google Lens are slated for their significant updates very soon, too. Other things will take a while, or never happen because they just don’t work as planned. The new Closed Caption system they demonstrated may never work well enough to actually put into production, and advances in healthcare will go through years of tests before anyone can decide if they are worth doing. AI and Machine Learning aren’t something you can build out overnight.
Google has plenty more to say about AI and how it can work with the devices we use every day. We’ll keep this page updated as google says more at Google I/O 2018.