Google’s homegrown smartphone hardly challenges the status quo all-screen, jumbo-sized designs from Apple and Samsung. On the glass back, there is a fingerprint reader and one camera; on the front, there is a notch in the screen to hide its front-facing cameras. The one flourish: a brightly painted power button.
But Google says it is not so much about the hardware — it is what the Pixel 3 can do. The Internet giant has made its own phone a vessel for its advances in artificial intelligence, which promises to personalize online experiences and save people extra seconds in a multitude of ways.
“The magic of this phone is on the inside,” Google’s senior vice president for hardware, Rick Osterloh, said in an interview. “AI really changes things quite a bit. It can have an understanding of you.”
AI can also require a lot of personal data. And Google’s hardware event, scheduled weeks ago, fell awkwardly one day after the company’s largest recent data-privacy scandal. On Monday, the company admitted it had exposed data on half a million accounts — and did not disclose it for months. On Tuesday, CEO Sundar Pichai did not make an appearance at his company’s event.
That defines the balancing act for Google, eager to finally make a dent in the premium phone market. Can customers trust even more of their lives in the hands of the data giant? And are its AI tricks useful enough to convince people to drop Samsung, or even Apple, for Google?
Rival Apple has played up its privacy-first practices and how it collects as little information about its customers as possible. Google remains, unapologetically, all about the data.
“The reason we have your data, in some cases, is because we are offering useful service. You trust us with your Gmail, and we keep that data private and secure. These are huge important issues we focus on,” Osterloh said. “Anyone who claims you don’t need data to offer a valuable service isn’t telling the full picture.”
The Pixel 3 packs in a host of new ideas about how AI can change the phone experience, from the practical to the pretty out there. Not all of its AI capabilities require sending data back to Google over the Internet, Osterloh said — in some cases, it is faster to do the analysis on the phone itself.
Foremost, there is the camera, which Google says uses AI to know a good photo when it sees it. The Pixel 3 starts shooting a second before you press the button and a second after — looking for the moments where everyone has their eyes open and facing the camera. If the Pixel 3 thinks it got a better shot than you, it will pop up as a suggestion.
Or, you can turn on a mode called Photobooth and let the camera decide on its own when to take the shots. (This uses a version of the software Google put into a stand-alone AI camera called Clips.) That is useful for selfies — you do not have to awkwardly press the shutter.
AI, embodied in Google’s talking Assistant, also promises to be more of a sidekick in the Pixel 3. That starts with answering the phone for you. When a call comes in, three buttons come up — answer, reject, or screen call. If you press that screen button, the Pixel’s AI will answer and start asking questions of the caller on your behalf.
On the Pixel 3 screen, you see an instant transcript of the caller’s responses so you can decide whether to pick up. I would not use it on mom, but it definitely might help with robo-calls.
Starting next year, Pixel phones will also be the first to get access to experimental — and controversial — Google software called Duplex. It uses a real-sounding AI voice to complete real-world tasks over the phone, like calling a restaurant to make a reservation. First previewed at Google’s I/O Conference over the summer, the technology sparked a debate about the ethics of computers pretending to be humans. (Google says Pixel phones will announce they are robots when they make calls.)