Each week David Souter comments on an important issue for APC members and others concerned about the Information Society. This week's blog post looks at the impact that digital assistants might have on our lives.
I wrote last week about new paradigms for the ICT sector – new ways in which ICT businesses can reach out for our attention, time and money. The trick for online business here doesn’t lie in offering new services alone. New interfaces between us and their technologies, if they’re successful, are an even better way to maximise commercial value.
We’ve seen the impact that new interfaces have over the years in ICTs. Apple’s graphical user interface and its Microsoft successor, Windows, for example. The World Wide Web. Smartphones and tablets. Today’s digital businesses are hoping that the ‘digital assistant’ is the smart new interface.
What is a digital assistant?
So what is meant by ‘digital assistant’? Some mix of hardware and software we can talk to that will do things when we ask: schedule meetings, for example, or remind us of them; turn on the TV; show us the way to Amarillo; choose a recipe for dinner.
There are already digital assistants out there, in one form or another. Lots of us have them on our phones and our computers, though we may not use them – Apple’s Siri, Microsoft’s Cortana or, in some ways, Google Maps. Some of us have bought into voice-activated home controllers like Amazon’s Echo, to manage those Internet of Things devices that are creeping into our domestic lives. These are immensely valuable, for examples, for those with disabilities.
People talk to pets that don’t understand them as they walk around their homes, so it’s not surprising that they quickly get accustomed to conversing with digital assistants (which answer back and can have virtual personalities).
The next steps
But digital businesses have bigger things in mind. I’ve attended two presentations lately on where they see this heading. Google’s Assistant, claimed Google at one of these, will not just recognise your needs, it will anticipate them; it won’t just converse with you, it will ‘personalise your relationship’.
Take an example. If I arrange to meet someone for lunch, my digital friend will notice, book the restaurant (and alert it to my eating preferences), order an Uber (or book the parking slot), commission the briefing paper that I’ll need and find someone else to pick up kids from school. Using algorithms, without much need to know on my part what it’s doing, it will make decisions for me.
A digital assistant’s like a housekeeper or a secretary, only not human. Surprise, surprise, digital businesses like to give them women’s names and voices.
What’s not to like?
I can see this being very popular. Voice-activation is more natural than a keyboard. Who wouldn’t want to save time and effort by getting the cloud to do things that would otherwise be inconvenient? In sci-fi fantasy, a robot; in the short term, a digibox somewhere in the corner of the room; in the longer term, perhaps an implant in the wrist – the digital assistant that’s always with us because it’s part of who we are.
I think digital assistants of some kind are likely to be commonplace in future. They’re going to be particularly valuable for those with disabilities. But, as we’re learning, like all new technologies, there’ll be consequences that we’ve not planned for and which, in time, we’ll wish we had. I’ll mention four.
We have been locked in here before
One reason that big digital companies like digital assistants is that they’ll lock us into their worlds much more strongly than we are today.
In the early days of the Internet, many people accessed it through what were called ‘walled gardens’. Using the Net then, before the modern Web, was difficult. Services like Compuserve and AOL offered one-stop shops with friendly interfaces that gave users what they knew they wanted without challenging their technical ability or exposing them to what they didn’t. Many people now use Facebook much like that.
Your future digital assistant – choose from those offered by, say, Apple, Microsoft, Amazon, Google, Facebook – will be just as powerful a one-stop shop as Compuserve or AOL. We’ll become dependent on it, and thereby dependent on the range of services its parent company provides. If it’s not using its own company’s services, it will require deep integration with those of partner businesses. Readers of last week’s post may recognise this as iDATE’s ‘club’ scenario.
What does this mean for competition?
I’ve written elsewhere about the tendency for online markets to consolidate. The larger online services (like Google and Facebook) become, the more data they acquire. Greater data volume improves the quality of outcomes offered by their algorithms, which in turn attracts more users. There are also economies of scale in data centres that should be taken in account.
This doesn’t necessarily lead to monopoly; more likely to a small group of powerful service providers that compete with one another and have the power to inhibit market entry – what economists call ‘oligopoly’.
That’s a problem in itself, but it’s compounded – from a competition point of view – when those businesses use market power to reach into adjacent markets. If my digital assistant books a restaurant, it will use an online restaurant booking platform. If it books a taxi, it will use Uber or another online taxi service. If there are competing providers for these services, which one will it use? That which its own platform already owns or has invested in, of course. Deep integration which will probably exclude small local services.
In traditional markets, regulators would intervene here to require more competition. In digital markets, that is much more difficult.
What happens to our data?
Thirdly, while our digital assistant’s serving us, we’re also serving it (or the data businesses enabling it). As we well know, those businesses make money from monetising what they know about us. That’s already much more knowledge than we might like. Imagine how much more again our favoured online service providers will know about us when almost everything we do passes through our/their digital assistant. And how easily different data sets can be combined to multiply the value of that information. Which can be traded on, to some extent, depending on data protection rules.
One danger in this is that someday – not now, perhaps, but someday – those data could be used to manipulate us rather than just to monetise us. If we trust our digital assistants to choose our restaurants, will we also come to let them choose our partners or how we cast our votes? After all, existing algorithms come close to this by pre-selecting options that they offer us to choose from (think TripAdvisor, Tinder, Facebook’s choice of news).
The second is in just how much gets known. Your digital assistant will know everything about you and your lifestyle: your daily habits, personal relationships, sexual orientation, health care problems, indiscretions. If it’s shared with other members of your household, it will know theirs too. How comfortable are you with that? How comfortable would you be if your Alexa shared information with your friend’s Alexa?
What does it do to us?
Lastly, it’s worth thinking more deeply about how digital assistants might change our lives in general.
On the plus side, they’ll undoubtedly be most convenient. Like PCs and mobile phones, our digital assistants will save us time and money. The better their algorithms become, the more they’ll do on our behalves. There are going to be many gains emerging from the loss of chores.
But there are also losses emerging from these gains. We’ll be ceding some of our autonomy for the sake of convenience. It may feel like we’re ceding it to our friend Alexa or Cortana, but in fact it’s their company that we’re empowering. We’ll make fewer independent choices about quite important aspects of our lives. We may lose the joy of discovering things for ourselves if the food we eat, the bands we listen to, the videos we watch on YouTube are chosen in response to what we know we like. There’s a risk we’ll know less about people who aren’t like us because our digital assistant knows that they’re not like us.
I’m not, in short, against digital assistants. For some people, especially those with disabilities, they’ll be vital. For others, too, they’ll have great value. I suspect that they’ll become pervasive in homes and offices, in richer countries, fairly soon. The question’s how to make sure they do things for us rather than deciding what we do.
Next week I’ll write about UNESCO’s search for indicators of ‘Internet universality’.
Image: Advertisement 1927 film Metropolis, on Archives New Zealand