Published in Jan-Feb 2018
Me: Alexa, what is the quickest way to get from Karachi to Lahore?
Amazon: Sorry I could not find what you are looking for.
Me: Alexa, what is the best way to get from Karachi to Lahore?
Amazon: Sorry, but the jury is still out on that one.
You will have spotted by now that I have acquired a new Amazon Echo for Christmas. Voice search is the new frontier and heavy hitters of big tech (Amazon with Echo and Apple with Siri) are busy boosting its benefits and battling it out to dominate. So, I thought I had better do some experimenting on behalf of Aurora readers.
Tech enthusiasts say that a combination of huge volumes of live data, combined with machine learning, will make voice search better and better over time and that the likes of Alexa will become something even more powerful and useful – personal assistants that we use to run our lives.
Let’s unpick this claim a bit.
Take my first question: “What is the quickest way to get from Karachi to Lahore?”
The answer is, of course by plane. I expect that in a few years’ time, Alexa will get this right – and will be able to give me a timetable, search for available seats and even make a booking. (Brand owners will have to track how voice search is different from typed search and make sure that they are discoverable – Search Engine Optimisation [SEO] will need to be evolved and updated).
Echo will develop its skills (to use Amazon’s word) at performing narrow tasks like telling the time, setting a timer, playing music, tuning into my favourite radio station, making purchases, suggesting replenishments for food I might have run out of and answering basic factual questions. It already does many of these things just fine. It even knows who was the first president of Pakistan. The Echo should be a hit with people who love quizzes.
All of this might make my life a bit more efficient. But will it make it better? Will it live up a grand ambition to be ‘my personal assistant’? Or is this language the kind of hyperbole that attends much big tech innovation?
Nobody quite knows how the AI that supports Alexa will develop – or what will be the confluence of technologies that enable breakthrough innovations in 10 years. For example, the way we use Facebook today (sharing images and films) was not imagined when it was invented; it is the product of a combination of innovations in devices, cloud computing and broadband. The distant future of technology is less knowable because of the unpredictable interaction between several enabling technologies.
Let’s take my second question: “What is the best way to get from Karachi to Lahore?”
More ambiguous. Do I mean ‘best’ in the sense of most pleasurable, giving me a view of the landscape, or most relaxing, or most likely to lead to chance encounters with interesting people? In fact, for all these reasons, I prefer (when possible) to travel by train than plane. Yet, I can’t see how Alexa can know this. I don’t send off the ‘data signals’ – a piece of jargon we used at Google – for Alexa to work this out.
Sure, I buy Lonely Planet guides on Amazon – so Alexa might assume that I like ‘culture’ (in the broadest sense) rather than lying on beaches. But nowhere in my e-mails or Facebook posts are there clues that I love train travel. Alexa could use ‘proxy data’ (the behaviour of people like me) to predict what I might like. Amazon already does this when you make a purchase – it tells you that the people who bought that item also bought X, Y and Z.
All of which is fine and dandy – and yet, another example of data being used for ‘relevance marketing’. But it stops way short of the insights into my tastes that will be needed if Alexa is to win an important place in my life.
There is another problem: Alexa can ‘scrape’ data that gives an insight into my past behaviours and interests. But what may strike my fancy in the future? In fact, I have no idea what that might be. Take the most recent issue of The Economist which carried articles about lifts for skyscrapers, the Icelandic language and the opium wars. When I picked up the issue, I did not know that I was going to find these topics fascinating. Alexa will struggle to pull off the same trick of serendipity and surprise. It is more likely to offer me a dulling feedback loop on my past interests and behaviours. Big data has yet to prove itself to be as culturally enriching as promised at the dawn of the digital revolution. Far from enabling an exciting mix of cultural discovery, it tends to serve up that which is most popular – an endless recycling of ‘greatest hits’ and ‘top 10’ clickbaits.
What about the distant future? Say 10 years hence. Nobody quite knows how the AI that supports Alexa will develop – or what will be the confluence of technologies that enable breakthrough innovations. For example, the way we use Facebook today (sharing images and films) was not imagined when it was invented; it is the product of a combination of innovations in devices, cloud computing and broadband. The distant future of technology is less knowable because of the unpredictable interaction between several enabling technologies.
Human nature however, is more constant and so I think we can see trouble ahead for big tech. The big issues are trust and our desire for control. There is a big difference between letting Google, Amazon or Facebook serve up messages that are likely to be relevant and allowing them to have access to so much of my data (such as web-searches, content of emails, purchases and location) that it can learn to be truly personal. I would need to trust them a lot to allow this – and the signs are that the trust in big tech is eroding – fast.
The year 2017 was when the scales started to fall from our eyes. In the early naïve days of the digital revolution, we happily shared our data (just click accept on those Terms and Conditions without reading them!) in return for free stuff. And in so doing, we have sleepwalked into a new type of enslavement that we are only now waking up to. Clues to this shift in culture are in the language – we talk now of scraping data and surveillance marketing. The phrase ‘big tech’ carries with it the Orwellian echo of Big Brother. When I joined Google, I had to do an online ethics course that warned against using the wrong language. So, phrases such as ‘crush the competition’ or ‘dominate the market’ were to be avoided – and above all, not written down, lest they fall into the hands of the trust busters to be used as evidence. Instead, we should talk about partnership and collaboration. All this is understandable in a legalistic age – and not necessarily proof of bad intentions. But the fact remains: big tech has really become huge tech or worse, ‘tech behemoths’. Since the days of Cicero in ancient Rome, we have always known that too much power corrupts as night follows day. It creeps up on you.
This is amply evidenced by the amount of money big techs now spend on government lobbying and lawyers.
Voices calling for the breakup of Big Tech emerged in 2017. A new book – Move Fast and Break Things by Jonathan Taplin – set out the agenda and then Scott Galloway, an influential US academic fleshed it in a series of lectures. (To see these, search for L2 on YouTube).
Techlash is now gathering pace. The Economist predicts it will remain in the public eye – “As worries about tech monopolies build, Silicon Valley’s secretive tycoons will have to persuade citizens in Europe and America that their firms should not be broken up or heavily-regulated.” (from The World in 2018) So, will Alexa become a valued personal assistant or an unwelcome Trojan horse in my home? Or just another device that will be forgotten and gather dust in a cupboard along with the Breville toaster? Time will tell.
Julian Saunders was Strategy Director, Ogilvy and Head of Strategy, McCann Erickson. He has worked on behaviour change campaigns for the UK Government and on innovation in The Zoo at Google. He blogs at www.joinedupthink.com