A Blog by Jonathan Low

 

Jun 21, 2012

The Discovery Economy: Your Voice is Your Passport


We are making the transition from voice activated to voice enabled.

The difference is subtle but important. Activation was revolutionary enough, though the underlying technology was based on applying principles already well understood. It was - and is - about access. Enabled means that voices, like fingerprints, eyeballs and other forms of biometric identification may now be able to unleash a more powerful set of interactions that will further the time/space/convenience continuum that rules commercial - and increasingly personal - life.

This latest set of innovations has been prompted by the success of Apple's Siri, the infamous iPhone 4 afterthought that turned out to be the star of that particular show. While initially derided, it has proven to be a handy tool. And it is a dramatic first step in the march from voice to text to understanding to meaning. Wisdom may be next, but in the interim, the commitment to this journey suggests that there is sufficient demand to make further positive developments likely. JL

Kit Eaton reports in Fast Company:
Captain Picard asked many things of the voice-controlled computer aboard his starship Enterprise--including "Tea, Earl Grey. Hot." Though your smartphone is unlikely to spit out a hot beverage, there is suddenly an urgent rush to add similar sort of voice-querying power to next-generation devices. But it's not all about the ease of control, or even hot refreshments. It's all about the future of search, which is to say--discovery of new things for you, the end user.

LG this week revealed Quick Voice, a personal voice-control assistant which features natural-language processing technology to let its smartphone users search, write emails and so on--with built-in support for interactivity with 11 apps at its start
It'll roll out in Korea at first on just a couple of devices, though LG hasn't said it will progress outside the country. But given that LG's also decided the tablet market isn't one it can win in, it seems likely that an international expansion of Quick Voice may be in the works to differentiate LG's smart handsets.

Samsung, market leader in Android smartphones, has been offering its own version of the same system, S-Voice, for a while now on its Galaxy devices--including its hottest new device, the S3. Though the system has not necessarily pleased reviewers, it does seem powerful on paper--offering the option to access it by saying "Hello Galaxy" (the equivalent of Star Trek's "Computer...?" request), launching apps, searching contacts, taking a photo and even unlocking the device.

Perhaps LG and Samsung are in a rush because new rumors suggest that Google is pushing aggressively ahead with its own voice-controlled assistant, set to be the natural successor to the slightly limited Android Voice Actions system. In a fabulously geeky nod, the tech is code-named Majel, after Majel Barrett-Roddenberry, who was both Gene Roddenberry's wife and the voice of that same intelligent computer aboard Enterprise. According to sources speaking to the Wall Street Journal, Majel was progressing quietly as usual, but has been given a powerful kick--and we may even seen her demonstrated at Google's I/O event on June 27th. While it's not known exactly what Majel can do, we can guess that, given Google's booming expertise in semantics, she will be as good as, if not better, at understanding user's natural language requests as any other of these services. Possibly even better, as Google has a vast database of language from the search queries billions of users have typed into its interface.

Or perhaps the impetus in this tech really comes from Apple. Because with Siri, no matter what its detractors say, Apple scored a direct hit with its iPhone 4S...and that was when Siri was merely a limited beta service that could only pull off a few actions, crack a joke or two and occasionally send SMSs to the wrong recipient (a personal pet peeve). Now Apple's freshly revealed that Siri will be getting a big power-boost in iOS 6, including an API so more third-party apps can interface with it, and Siri will be coming to the iPad--the market-dominating must-have tablet. Given how PR-friendly Siri is, and that she's even making headlines for moving onto a safety-conscious "eyes free" button on certain marques of car, Apple's rivals are likely worrying that Siri will carry the next iPhone and iPad and other devices to new heights of attractiveness to consumers.

So what's all this voice-activated business about? It's about doing much more than delivering your request. Picard never got (well, hardly ever) anything other than his cup of Earl Grey. These companies are trying to do much more than this, and suggest new things to you. But why?

When you access these systems, you need to think about who gets access to your vocal request. It's not like tapping a search term into a query bar on your iPhone or Android phone, as we all do every day. Instead, when you speak into, say, Siri, it's Apple's servers that get first stab at the data. This has even caused a security flap recently with critics crying "privacy invasion," despite the fact that it's exactly this sort of crowdsourced collection of voice samples that enables our impressive 21st century voice-recognition powers. But it also means Apple gets to create a new database of search terms--exactly equivalent to the one Google uses to power its billion dollar-making user database. Apple's system even smartly works out whether you're looking for a piece of hard data (which it thus passess off to Wolfram Alpha) or other requests.

That's why Google's worried, and that's why it's making Majel. It's also partly why Samsung and LG are in this game. Controlling the flow of a user's search query means you control the power because you can accumulate customer intelligence, and thus do things like sell targeted advertising at them. Ask Google--it's in hot water for funneling user searches toward its own services over those of rivals.

And ultimately this is great for us as consumers of voice-controlled phones. You may not use them now, and you may pooh-pooh the limited Siri. But when Siri gets cleverer, and when systems like Majel come online you will use them because they're convenient, but also because they can deliver relevant information to you more quickly, and indeed improve your discovery of information. Think about Siri--she can tell you if you need an umbrella today or not, given knowledge about where you are, an understanding of the semantic context of your request, and direct access to a database. One day soon, she, or one of her peers, will be able to answer questions like "What aircraft is that flying overhead?" and, critically, "Can you suggest a good wine bar in Paris near my hotel that I've not tried before?"

0 comments:

Post a Comment