Jo Bowman
Might your colleagues one day be robots, your research merely simulated by some super-intelligent machine? Professor Nuria Agell of ESADE Business School in Barcelona thinks that, one day, the truth may be stranger than science fiction, writes Jo Bowman
At first, it sounds a little far fetched, but a decade ago, much of today’s mobile technology would have sounded like it came from an episode of Star Trek, too. So consider a moment the idea that, as well as crunching the numbers that quantitative research generates, a computer might one day be able to generate original insights from that data. It might be able to link today’s data with research done years ago to draw another conclusion, or remember a magazine article it read that was relevant and incorporate that, too. As well as processing numbers, it could also process facial expressions, locations, eye movements and comments elicited and overheard. And it might even come up with the next big breakthrough product idea – or enough of an idea to inspire a human being to perfect it. This is the kind of world predicted by artificial intelligence developers, and for market researchers, it is, potentially, an enormous aid to finding better insights and creating better, more targeted products, services and communications. There’s also a chance it could give researchers a strong push along the road to obsolescence.
What counts as intelligence is a sizeable philosophical question in itself; what constitutes artificial intelligence is an even more complex subject, one that, across the scientific community, is the subject of lively debate.
For Nuria Agell, artificial intelligence can span a range of applications, from a book-recommendation algorithm on Amazon.com, to a robotic companion that “knows” just the right thing to say to cheer you up after a bad day at work.
“Artificial Intelligence is a branch of computer science – but there are also people collaborating from cognitive science and mathematics and engineering – that aims to build or construct machines, systems, software or objects with the capacity of reasoning similar to human beings,” she says.
Algorithms that can suggest products or music you might like, based on your browsing preferences or purchase history, are already commonplace, replicating the role that a friendly, knowledgeable record shop owner would have had in days gone by – except with no lapses of memory, no bad moods and no need for a lunch break. This kind of ability, without human failings, has obvious appeal to business. “Humans are good at that, but sometimes our reasoning or what we do depends on whether we’re feeling more or less optimistic and whether we’re feeling tired and can’t bring all the information we have to mind all the time,” Agell notes.
Taking the plunge
But it’s beyond this, and into the realms of consumer insight and concept development, into what’s already known in some circles as business intelligence, where artificial intelligence promises to shine: learning by experience, searching and planning, reading nuances in the way people express themselves, and working on several projects or ideas at once. The common-sense element that human beings can inject into a process remains the X factor that’s at the heart of generating artificial intelligence.
“One of the challenges of artificial intelligence is to be able to use common sense that human beings have, so to be able to compute without using numbers,” Agell says. “Computing with words, and orders of magnitude instead of specific values, is one of the main differences between human beings and sophisticated pieces of computing. If you see someone about to jump from a window in a building near your office, you, as a human being, don’t need to know how many metres he’s going to fall. You can push the alarm button because you know someone’s about to commit suicide, and you do that without computing all the numbers.” Businesses need systems that can respond in the same way, sensing the early-warning signs of a crisis long before sales fall through the floor. “If we get the capability of reasoning but all the capabilities of a machine, with no memory loss and no tiredness problems, that can be very, very useful.”
Artificial intelligence promises, Agell believes, to provide more accurate and more useful analysis of consumer behaviour, picking up and explaining apparent anomalies, and then suggesting ways that businesses can use what seem like quirks and unpredictability in what people do and buy to present them with just the thing – a brand, product, offer or message – that’s right for them at that moment. Segmentation will change to account for the fact that people behave differently and identify with different groups and brands depending on what they’re doing, who they’re with, even the day of the week. “We want to catch the ambiguity we have when you go shopping and decide. Maybe on a Monday I act as a working woman, but when I go in the same shop on a Saturday, I’m acting as a mother or a grandmother.” Research that takes into account the context of a person’s decisions will make segmentation more realistic, she says. “Sometimes you can be in a group of people and always buy a certain type of product, it’s what belongs in that group; but in another group, where you might behave in different way, I can offer you something else.”
Ask an expert
Another application of AI in business that ESADE has been experimenting with is in tapping some of the specialist knowledge that experts in a particular field may have, loading that into a machine, and then using that to have the machine generate new ideas. While a sommelier might not be able to articulate or quantify why he’s recommending a certain pairing of wine and a meal, and an advertising copywriter doesn’t know where her great slogan idea for running shoes came from, Agell thinks it’s possible that, by capturing some of their knowledge and teaching a machine to “think” in the same way, some of their reasoning can be imitated in a way that allows that sort of insight to be replicated, or at least generate new, informed inspiration for the experts. “People like product designers or chocolatiers or a perfumist, when they’re thinking about what they’re going to create next, sometimes it would be interesting to have a machine that helps them,” she says. “There’s a branch of artificial intelligence that tries to capture this expert knowledge that sometimes is very complicated. These people have very privileged senses, and we try to capture in a passive way the naïve, abstract reasoning they use . . . and use what they already think to recommend new things, in the same way as AI can recommend new things for consumers to buy.”
In a project late last year with the renowned Spanish chocolatier Oriol Balaguer, researchers introduced information provided by an expert chocolatier to a computer that was “taught” to use that knowledge and recommend new combinations of chocolate and fruit that were likely to be popular with consumers. Samples were created according to the machine’s suggestions and taste-tested among a small group of young chocolatiers. Their responses were fed into another AI-supported computer system that analysed their responses to determine the most-liked product out of the handful being tested. Agell says the small focus group was faster and more effective than it might otherwise have been, because artificial intelligence allowed participants’ opinions to be given in a more flexible, natural way, with non-numerical scores and words. A new fruity chocolate cake is in development as a result of the project.
With an ability to collate results and generate insight from a huge range of sources, AI promises to make market research more efficient and more effective, Agell says. Artificial intelligence can extract meaning from the wealth of information – quantitative and qualitative – that businesses already have within their organisation, linking the results of everything from eye-tracking studies to sales data and web site click-through rates, and then applying human-style common sense reasoning to draw conclusions.
“Creation and innovation are a little more difficult, because the way of capturing information from human experts is more difficult, but I believe that it’s also a very important chapter, to put your expertise in a machine,” she notes.
Only human
Not everyone, of course, is delighted by the idea that their years of experience and unique perspective and inspiration could be turned into a computer program. There have been what Agell calls psychological barriers, reluctance to share expertise. “Sometimes they’re sceptical, because they think that if they give this information to someone else they can substitute them, but it’s really just to help themselves.” Will artificial intelligence ultimately put everyone out of a job? “Absolutely not.” On the contrary, Agell insists that artificial intelligence is a way of strengthening people’s performance in their jobs, helping them make their work more valuable. “This is the way of creating new jobs, because the technology allows us to think of new ways of working, new combinations, but all this technology has to be close to people – the experts or the designers or shoppers who need help in their job. It has to have a human who has to say if it’s going to work or not.” And it’s a human who will carry the can – or take the credit – for what artificial intelligence may recommend. “The machine has no responsibility,” Agell points out. “If your boss asks you to define a new concept, if the machine decides, you rely on the machine and it doesn’t go well, it’s you who’s going to lose your job, not the machine.”
Differentiating between human and computer capability is already becoming tricky. Websites that use CAPTCHA codes to try to ensure that real people are using their sites are having to make the tests increasingly difficult, as web bots are becoming more adept at ‘reading’ them. In some cases the result has been CAPTCHAs that are too difficult for humans to decipher, and mini-games from sites like areyouahuman.com have begun to emerge as replacements.
Agell expects significant advances in the use of artificial intelligence over the next decade or so, not just in business but in our everyday lives. “There’s a lot of applications with small robots designed for the elderly, to be a companion, like a pet. There can be intelligent companions that play with kids in hospital when they feel alone – a toy or robot that can capture the emotions of the human being that’s with them. It gets used to you, and when you’re a little bit down, it acts in a different way than when you’re happy. I expect lots of developments in the next 10 to 20 years.” Perhaps there’ll be a companion robot to cheer you up when your role in the office becomes obsolete.
Nuria Agell is a professor at ESADE Business School, Ramon Llull University, Spain
1 comment
Firstly, we should forget the vague and silly word “intelligence”
The feature which is remarkably highly developed in the human is imagination.
Imagination being quite clearly definable as the ability to form, store and morph neural representations of the external world. Greatly enhanced in our species by the transfer and external storage of imagination which we call language.
But to look to the main issue:
While would be a very time before any individual computer can surpass the cognitive flexibility of the human mind, the current integration of myriad such systems will indubitably reach this stage within a few decades.
Not, as we may dearly wish to believe, there by the specific design or for the benefit of humankind.. Rather it is the culmination of a process that has been brewing within the collective imagination of our species for around 2.5 million years. And, while we undoubtedly benefit substantially from some of its spin-off, we must remember that we are but a very small cog in a vast evolutionary machine. Humiliating as that may be.
Our aanthropocentric conceits tend to make us quite oblivious to the remarkable developments taking place right under our noses.
The construction of a “brain” that will soon equal and then surpass that typical of our species has for long been a work in progress. Not as a result of any deliberate human “design” but rather as the result of an autonomous evolutionary process that can be seen to have run its exponential course since humankind acquired the ability to share imagination, which we know as language.
It can already be observed as a a work-in-progress. And effectively evolving by a process of self-assembly. You may have noticed that we are increasingly, in a sense, “enslaved” by our PCs, mobile phones, their apps and many other trappings of the increasingly cloudy net and its ever-accumulating insidious peripherals. We are already largely dependent upon it for our commerce and industry and there is no turning back. What we perceive as a tool is well on its way to becoming an agent.
Consider this:
There are at present an estimated 2 Billion internet users. There are an estimated 13 Billion neurons in the human brain. On this basis for approximation the Internet is even now only one order of magnitude below the human brain and its growth is exponential.
That is a simplification, of course. For example: Not all users have their own computer. So perhaps we could reduce that, say, tenfold. The number of switching units, transistors, if you wish, contained by all the computers connecting to the Internet and which are more analogous to individual neurons is many orders of magnitude greater than 2 Billion. Then again, this is compensated for to some extent by the fact that neurons do not appear to be binary switching devices but can adopt multiple states.
Without even crunching the numbers, we see that we must take seriously the possibility that even the present Internet may well be comparable to a human brain in processing power. And, of course, the degree of interconnection and cross-linking of networks within networks is also growing rapidly.
The culmination of this exponential growth corresponds to the event that transhumanists inappropriately call “The Singularity” but is more properly regarded as a phase transition of the overall “life” process.
The broad evolutionary model that supports this contention is outlined very informally in “The Goldilocks Effect: What Has Serendipity Ever Done For Us?” , a free download in e-book formats from the “Unusual Perspectives” website