Voice recognition and the future of search
Science fiction has done loads to shape our perception of how the future will be. And despite the wealth of exciting and innovative technology now out our fingertips, one common complaint these days is “it’s the 21st century, where’s my flying car/jetpack?” But for me, one striking omission in our repertoire of gadgets is the ability to use speech to directly issue commands to technology.
Great strides have been made in this area lately, as demonstrated by the likes of Apple’s Siri and Google’s advanced voice search application and the function seems to be inexorably linked to the future of search. But what powers this enigmatic technology, why has it taken so long to arrive and how will it affect the way we interact with the web in the future?
• Why Do We Want It?
Voice recognition is an area with many applications and has been sought after for more than a century. Alexander Graham Bell and his contemporaries arguably made the first forays into the field with research that eventually led to the Dictaphone and telephone.
Bell aimed to create a device that could signify sound to the deaf and hard of hearing and many of the obvious applications of voice recognition technology are as an aid for the physically disabled.
However, in today’s ever-more-connected world, accurate voice recognition can have a profound effect on the way in which we interact with technology and the web. This is especially true when taking in to account just how attached we are to our mobile computing devices.
But don’t just take my word for it. In a recent media presentation Google revealed how much weight it’s placing on the field.
“If we are going to build the search of the future, we will have to solve difficult technology issues like speech recognition and natural language,” it said.
• What’s The Problem?
With all the interest, you’d think voice recognition (also called automatic speech recognition or automatic speech understanding) would have shown up a lot sooner. So what’s the hold up?
There’s an array of problems when it comes to teaching computers to do something that comes so naturally to us. For a start, there’s the issue of dialects – and not just regional ones. It turns out that we have different social dialects depending on who we’re talking to. And that’s just the tip of the iceberg.
Some other stumbling blocks for speech recognition include slang, background noise, the disparity between spoken and written language, the length left between words (known as word boundary ambiguity), the recording hardware used, variability of speech within a single speaker, the sex of a speaker and the speed at which we talk.
Voice recognition technology has ploughed valiantly forward over the years, but only started creeping into the mainstream in the 1990s, mainly due to the computing power required. One of the most recognisable early applications took place in call centres. The AT&T Operator line was a pioneer in this field that helped these facilities route calls to the right place via automated voice recognition.
However, this unglamorous application was a far cry from the human-machine interaction made famous by the likes of 2001: A Space Odyssey and Star Wars. In fact, it was only in the late 1990s that we saw the introduction of software that could be used on your average PC.
The 21st century has seen major advances in the field. Voice recognition applications no longer need to be arduously trained to attune to a specific user and in addition to Google and Apple’s offerings, there are a range of apps that enable relatively straightforward tasks to be completed via a voice interaction-powered assistant.
• The Future
The repertoire of tasks that our mobile voice recognition assistants can perform is only set to grow and search will remain one of their most important functions. The growth of semantic search, which would enable software to better-understand the meaning behind queries, could have wide-ranging effects on the field.
Imagine, for instance, your personal assistant being able to not only find you a top-rated, local take-away but actually order you a meal. Similarly, while you’re currently able to find a taxi with Siri and co, in the future, you’ll be able to have them autonomously book one for you.
“Our young children and grandchildren will think it is completely natural to talk to machines,” Microsoft’s Eric Horvitz told the New York Times.
Microsoft has been one of the forerunners in the field and the coming years could see the company break down language barriers. Back in November, Rick Rashid demonstrated a really impressive real-time voice translator on behalf of the firm, which could open new realms of possibilities for the field.
Voice interaction is an arguably inevitable evolution of our relationship with computers. However, despite all the recent advances in the field there are still a lot of creases to be ironed out and unclear implications for the technology.
One such issue is how the user will interact with search. Assuming we one day get to a point where voice recognition-powered assistants are sophisticated to the point of being ubiquitous, how will information be imparted?
This could be fairly straightforward in cases where the query has a clear-cut answer. For instance, you could ask “when is Easter this year?” and receive a concrete date. However, when asking something like “what were the causes of the American civil war?” how much detail would be required? Similarly, how will search engines rank results and monetise a service where answers are simply spat out as needed?
Such problems are by no means insurmountable, but it’ll be really interesting to see exactly how they’re overcome.
Voice recognition and interaction are genies that are firmly out of the bottle and it’s hard to envision a more intuitive way of interacting with our gadgets. And when considered alongside exciting augmented reality products like Google’s Project Glass, it’s easy to imagine a future where technology takes a less visible but increasingly important role in our lives.
If you’ve any thoughts on the future of voice technology, be sure to hit me up in the comments and if you’ve liked this post, don’t forget to share on your social media of choice.
Images used by permission under the Creative Commons Licence courtesy of Custard Media.