Until very recently, you might have been forgiven for thinking that proper voice recognition was never going to arrive. Decades of experience with unusably dreadful in-car systems has equipped many of us with a healthy scepticism toward the technology.
But things seem to be improving rapidly. Thanks to the exploitation of ‘big’ data, and advances in machine learning, the technology has finally reached a tipping point where voice commands are, more often than not, understood. So why is this such good news?
Issuing voice commands is effortless. No longer do you have wield a mouse, keyboard or touchscreen in order to buy your favourite cereal – now all it takes is a sentence.
This is obviously fantastic news for online retailers. In fact, Amazon are so keen for us to place orders in this way that they’ve released the code for their Alexa Voice Service into the public domain, encouraging hobbyist inventors and smaller manufacturers to incorporate the voice-command functions everywhere.
Voice search is available even when you’re busy doing other things, or when your hands are occupied. So, you could be driving to and from work while having a conversation with your voice assistant, or asking for DIY instructions while you’re underneath a half-assembled wardrobe with both hands occupied.
Another considerable plus-point of spoken commands is that virtually all of us know how to issue them. Not only does it take time to click a mouse, scroll up and down a page, and navigate a social-media interface, but it also takes time to learn how to do such things. Not so with voice-recognition, which poses no learning curve to navigate. If you’ve ever seen an older person struggle to operate their mobile phone, then you might have an idea of the problem this solves.
One group that stands to benefit from this are elderly and disabled people who spend much of their time housebound. Voice search will allow them to easily call for help, and easily communicate with the outside world.
In this area, our spoken language, cumbersome and illogical as it often is, is considered the ideal to which machines should aspire. Thus search engines have learned to respond to sensible questions like “How do I strip wallpaper?” rather than demented streams of words like “instructions wallpaper simple stripping”. Indeed, instances of the words ‘where’ and ‘when’ have become enormously more popular in recent years.
Thus spread of voice search naturally has considerable SEO implications. The days of jamming keywords into every available sentence are gone. Instead, search engines will reward the more conversational tone in which search engine questions are often expressed. A subheadings phrased as a common question is a great way to draw traffic. For this reason, companies that don’t have an FAQ page are missing out on a straightforward path to that ‘snippet’ real-estate at the top of Google question searches.
A world where everyone gets their information using voice commands rather than screens is an existential threat for the modern advertising industry. While advertising banners around the sides of an article you’re reading might be unobtrusive, putting the same information into an audio advert that interrupts your experience certainly is. While it’s possible to display content and banner advertisements at the same time, it’s impossible to play their audio equivalents simultaneously.
Of course, the information that comes back from a voice command will often include detailed instructions, incorporating visuals as well as audio. For example, if you’re in the kitchen and you need to clarify a recipe for banana bread, having a visual demonstration appear on a nearby display is always going to be more convenient.
Happily, Google’s smart display platform and Amazon’s Echo Show open the door for integrated solutions, and most modern smart televisions already come with microphones installed. Thus we should expect screens to stick around even in a world where voice assistants answer our questions directly.
It’s easy to see where this technology could lead, if current trends continue. The future is one in which machines won’t just listen to what we have to say. They’ll also be able to confidently talk back to us, clarifying details and asking questions of their own.
The world looked on in horror and astonishment in May as Google CEO Sundar Pichai demonstrated an AI that could not only hold a conversation with a total stranger, but do so more convincingly than many humans could. Add to this the advent of ‘deep fake’ face-swapping technology and you have a formula that could render us totally unable to discern truth from fiction. This presents the possibility for all manner of online fraud. If cold-callers asking for your bank details are a problem now, just imagine if those crooks were able to assume the face and voice of your own children.
The more immediate future is one where businesses should aspire toward a more conversational tone for their online content. With younger generations particularly eager to adopt voice-search technology, ranking well on search engines may well depend on speaking their language.