Dr. Chris Brauer's predictions for the 'transparent' future of artificial intelligence
Ignoring the rise of virtual assistants now would be akin to writing off Google as a niche research project 15 years ago, believes Goldsmiths' Dr. Chris Brauer. Emily Hare spoke with Brauer about the implications of artificial intelligence for the marketing industry in issue 42 of Contagious.
Brauer will be speaking on the topic at Most Contagious 2015, December 9 in London. Tickets available at mostcontagious.com.
Chris Brauer’s focus on the future has led him from wearables and data tracking to driverless cars, and now virtual assistants, starting with the likes of Siri and Google. Now, Brauer believes virtual assistants (VAs) will take a variety of forms, from robots right through to tiny pieces of nanotechnology. They will be on hand to provide us with all kinds of pre-emptive, contextual information, from what streets to avoid to what not to buy. The founder of the Centre for Creative and Social Technologies and director of innovation at the Institute of Management Studies at Goldsmiths, University of London, shares why he believes virtual assistants will become the primary way that we access information online and how brands can prepare for that scenario.
Contagious: What can we expect from virtual assistants?
Dr. Chris Brauer: They will effectively be like an angel in the cloud, pulled down into different environments. So your VA will exist in your vehicle, home, workplace, or while you go out running. But what form it takes will modify depending on the circumstance or the environment that you are in.
It also learns about you in very specific ways, from your movements, your interactions, your transactions. Is it your butler? Is it your servant? Is it your friend? Is it your master? Does it constrain your financial transactions because you have established some sort of rigid financial goal or because you are trying to reduce debt? A virtual assistant does not need to be screen-oriented in the future. It will be fluidly connected to human experience, rather than pulling a small computer out and interacting within a screen-based format. There is enormous potential in nanotechnology that allows us to create smaller and more powerful devices. [For example] there is a strong logic to having something in your ear that you can interact with. So if something is going to tell you to eat a banana, it makes a lot more sense that it is integrated into your experience rather than vibrating in your pocket or requiring you to activate something to access the information.
Do we decide whether our VA operates as a butler or a master?
A lot will depend on circumstance. Some people will prefer to establish the terms under which the VA interacts. And a lot is going to depend upon how design evolves and who is influential in its development. That is why all the major tech companies are significantly investing and moving aggressively towards the development of the platform. I think consumers will choose the services that the VA supplies and the way in which they interact with it.
Your research into VAs stems from your work into wearable technology. Why is that important?
Wearables don’t just augment our knowledge of the world around us, they offer us the opportunity to learn more about ourselves. In many ways, we have a limited understanding of what makes us tick and the conditions under which we are happy, or making money, or being productive, or emotionally moved and how these things correlate to various situations. Wearables make visible that which isn’t immediately apparent, giving us an opportunity to use that to support decision-making.
You are using your VA precisely because it has all the capabilities that a machine can bring and a human being can’t.
Your Human Cloud at Work project with Mindshare looked at how wearable data can be used to support decision-making in the workplace. What did you learn?
There are organisations where people are already being measured quantitatively in their performance. But knowledge-based businesses such as advertising often think that there isn’t a clear way to evaluate performance in that kind of way.
A football team would use real-time data from the performance of their players on and off the field to evaluate who is best to put onto the pitch. This is a crucial competitive advantage in sports, but hasn’t moved into a lot of other industries yet. If you are an agency bidding for a large contract you have to decide which members of your team are performing at the highest level or which are emotionally, psychologically, physically optimised to perform. An account director might decide who to send into a pitch meeting based on data from a wearable – not exclusively, but as an input. Individuals will also be able to take the data that they are getting from these devices and put together their own rationale, building a biometric CV. They can share that information with others and make broader changes not just to their lifestyle, but also to the work environment and their home.
There’s been a lot of discussion around machines passing the Turing test. Is that actually important?
The Turing test asks whether you can tell the difference between the machine and a human. You are using your VA precisely because it is a machine and has all of the capabilities that a machine can bring and a human being can’t. You will retain for yourself the parts of your life that are the domain of the human being – our free will and our decision-making. Superior VAs are so near to entering the marketplace and will become the dominant platform for accessing knowledge and information because they are explicitly a machine and present themselves as such.
If you look at mobile phones, people are already linked to that device in a way that far exceeds its functionality. A VA is nothing more than a smaller, faster and more efficient, more interactive version of that same concept. They are coming very quickly and brands and organisations that aren’t prepared will be caught off-guard in the same way as you would have been if in 2000 you looked at Google and said, ‘Well, that’s an interesting research project at Stanford.’
What should brands be doing to prepare for that kind of scenario?
It starts with transparency. I think the era of convoluted interactions with consumers is going to be over. The VA will be able to see through that kind of gaming of the consumer experience, so you are going to have to be straightforward: what is it? What services do you offer? Why are they useful? What do other people have to say about it? All the things brands are doing now will be important because the VA will look across all that information to assess what is the most useful resource. It’s time to think about how you fit into the next generation of the internet that is formed by this VA. That is what we are working on with Mindshare and its clients. Each brand’s conditions and the way in which it may be able to leverage the possibilities of these technologies will be different.
So marketing won’t be enough?
No, but marketing your credibility can be a valuable asset because you are marketing not just for the person but to the VA. So you start to market your transparency. If the VA or the VA ecosystem trusts brands in certain regards, then branded modules emerge. If you are dealing with your cholesterol, Flora is at the forefront of those interactions, making investments in technology that allows you to live a lower cholesterol lifestyle. Therefore why doesn’t it just position itself as a service which does exactly that? Yes, it sells margarine, but it primarily also sells a low cholesterol lifestyle and that is something that can integrate well into a VA environment.
As an early indicator of this, we’ve seen brands like Lowe’s and Bank of Tokyo Mitsubishi, UFJ working with AI in retail environments. What are the opportunities for retailers?
Each robot learns from its interactions and receives information back from which the other robots can learn.
The collective intelligence of these services will have a marked impact and that is why big retailers have a unique opportunity. They can roll these robots out across a global network of outlets and learn about cultural differences, purchasing behaviour, behavioural economics and context, and feed all that information back, collectively optimising as they deal individually with each situation.
Trust is going to be a key factor. Tell us about trust dynamics, particularly in relation to brands.
Say you go into a pharmacy and look at a branded version of an off-the-shelf painkiller and a generic house brand. The difference in price is quite dramatic. Behavioural economics has long exploited the opportunity for the branded one to position itself as offering something more, but when you look at the ingredients, they are identical.
The justification has been the way in which the products have been packaged, sold and marketed, but also how that has accessed the irrational tendencies of the consumer. Now the VA will know that those two products are identical in terms of their make-up. The VA doesn’t have an irrational brain or consumer behaviour black spots. It can’t be exploited in the same way as a human being. Will I trust my VA to tell me which painkiller to buy? Absolutely.
There are vast swathes of our decision-making where we have too many choices, too many things that we have to engage with and having a virtual assistant in place that can make all those decisions will be a useful functionality. We are talking about a hyper-rational machine and needing to develop strategies for engaging with that.
How do you see the uptake of VAs developing?
It is going to be very difficult to say, ‘I don’t really want the VA service,’ because you’ll be facing a situation where if you don’t let your VA monitor your kids’ behaviour online and your kids get bullied or groomed or exposed to something that you don’t want them exposed to, then what kind of parenting model is that? It is no different than going abroad and saying ‘turn off my fraud algorithm [designed to spot credit card misuse], I will trust my own intuition as to whether or not I am in that risk in any of these transactions.’ There are a lot of things where machine intelligence is better than human beings. Those are the early spaces that the VA is going to fill. Then, slowly but surely, it will simultaneously be a nanny, a friend, a servant, a master and whatever else you need it to be.
Tell me a bit more about your partnership with Mindshare? How are you working with it? What is the media agency looking to deliver to its clients?
We can see a future vision of the virtual assistant three to four years down the road, so where are the brands going to fit in? Our new project, Shift 2015, is commercially focused, where we are working together to understand the advertising, marketing and media-buying opportunities for brands in wearables and virtual assistants.
The next thing is to run a bunch of experiments with consumers, working co-operatively with them to understand where this fits in so that we can then answer some of the questions that you have been asking more clearly. It has been a very fruitful way to interact and it offers a different perspective to an independent academic research unit. And, as a corporate client, Mindshare can develop strategies that give it a competitive advantage in this area. Our hope is that we will be in a position whereby we understand the space and we also understand the specific opportunities that are available.
Are we ready for the changes that VAs will bring?
Everywhere I go I try to talk to people about that and about how they can get themselves in a position where they can feel more comfortable with that line of flight. If you are doing a five-year plan for an organisation, you are talking about a VA that in five years’ time is six or seven times as powerful as it is today. If you are putting in place a strategy for the middle of 2016, and we are at the beginning of 2015, then the technologies available will be twice as powerful as they are today. Your strategy should be oriented towards that.