At the Google I/O conference, Sundar Pichai, CEO of the company, revealed what perhaps will be the killer feature for Google’s Assistant aspirations. In a series of real phone calls, Google Duplex was able to interact with real humans to book appointments and get information. Duplex’s ability to simulate a real human was stunning: Not only can it understand spoken English with its insouciant disrespect for grammar, but it can also simulate the conversational tics natural to human speech across cultures. This is yet another reminder of how close we are to a world driven by AI.

As soon as I saw this, I immediately wanted it. Imagine how much easier and convenient my life could be if I had a real AI assistant making calls and booking things for me (big time sucks, especially when you have little of it in the first place). I mean, yeah, I can get my groceries delivered to my door in under an hour already, but think about making reservations at restaurants, perfunctory and repetitive conversations with car and cab drivers (“Where are you?” “At the place I said in the app when I booked the car…” “Oh, I’m at the roundabout, can you come over this way?” “I’d really rather not.”), and booking appointments with offline services… or even friends and colleagues!

Alas, China is not as voice-driven, and the development of consumer products has been lackluster. Sure, B, A, T, and even J, and X all have their own speakers… but does anyone actually use them? Do they even work? In our testing at TechNode, the results have not been satisfactory. Perhaps it’s my accent? Or maybe my tones aren’t right? But then again, even my accent is more standard (标准普通话) than most Chinese. According to a 2014 report by the Ministry of Education, roughly 7% of Chinese people can speak standard Mandarin smoothly. Compare this to the US where an overwhelming majority of people speak with a Standard American accent. Accents are perhaps the most challenging part of voice recognition and this is doubly so in China, a country with 100s of local languages, some of which are mutually unintelligible (traveling outside of Beijing can be a bit harrowing since my usually strong language abilities fall off a cliff when confronted by an unknown accent).

And it’s not just voice that makes something like Duplex so difficult in China. It’s also the fundamental difference between how we interact with each other, and the mobile phone’s role in mediating relationships. China isn’t very innovative (there, I’ve said it), at least in the technical sense. Much of the gains in China’s consumer economy, whether that’s efficiency, convenience, or user experience, have come from using the internet and mobile phones to bring traditional forms of commerce online. Dazhong Dianping, Didi, Taobao, JD, and even bike rental companies have all been able to grow quickly by putting traditionally clunky, slow, or costly transactions onto the mobile phone. It used to be that calling a business or showing up were the only ways to get things done; these days, calling is almost ineffectual as booking and queueing are all done either online or face-to-face. Even then, you’re not going to call your favorite hairdresser… you’ll send them a message on WeChat to see when they’re free.

In the US, the major players are all trying to get into your home. Amazon has a huge advantage as a first mover, but Google could perhaps have better technology. Those are the only two real players (don’t get me started on Siri), the ones that have near monopolies on how information and products are served. Across the Pacific, the competition space is quite different. Yes, you do have the 1st gen majors of BAT, but the 2nd gen are already making waves and spreading into other waters. The threat of disruption from rivals and upstarts is constant.

As a consumer, choosing between Android and Apple on mobile is relatively easy, choosing between Google and Amazon in the home is fairly easy as well… but how do you choose between an Alibaba speaker that doesn’t let you interact with JD or WeChat or a Tencent speaker that doesn’t allow you to buy stuff on Taobao? The cutthroat environment and anti-competitive moat building has created an extremely fragmented market for online services. The only real option is to go with Xiaomi, an independent making the only truly smart home ecosystem of devices and services. Even then, its still limited to playing music, changing the volume, checking basic information, and interacting with other Xiaomi devices.

There’s a cultural reason, too, that something like Duplex isn’t in China yet, if ever: fraud. Problematic for centuries, Chinese in my experience are deft at finding loopholes and exploiting unintended consequences. It’s only with national ID cards with embedded RFID chips that we can finally be assured of getting a train ticket without resorting to a “friend who has a friend.” Official receipts keep getting more and more complicated to prevent forgery. Those same ID cards are now used to prevent impersonation and identity theft to a degree only dreamed of in the West, with their qualms of authoritarianism and queasiness with increasing lack of privacy. Imagine the levels of fraudulent transactions and general mayhem that responsive voice-driven AI could cause if allowed to run wild in a culture with a long history of taking advantage of the bigger fool. This, I don’t imagine, is something the government would tolerate for very long if it even is allowed to exist in the first place.

I’ve certainly grown accustomed to living in China and, in some ways, my life has gotten better and easier since I first moved here, but sometimes I do wish I could have the best of both worlds.

John Artman is the Editor in Chief for TechNode, the leading English information source for news and insight into China’s tech and startups, and co-host of the China Tech Talk podcast, a regular discussion...

Leave a comment

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.