"If a machine can think, it might think more intelligently than we do, and then where should we be? Even if we could keep the machines in a subservient position, for instance by turning off the power at strategic moments, we should, as a species, feel greatly humbled.” – Alan Turing, the father of modern computing, in a 1951 talk on BBC Radio
IBM’s artificial intelligence (AI) platform, Watson, is loquacious; it can tell jokes, answer questions and write songs. Google’s AI can now read lips better than a professional and can master video games within hours. MIT’s AI can predict action on video two seconds before it begins. Tesla’s AI powers the company’s innovative self-driving car. All seem to propel us closer to Turing’s world of machines with more intelligence than humans.
If Turing’s words now ring true, should we feel humbled or anxious? For many marketers, the anxiety and existential fear has given way to hope and excitement for a new tomorrow.
“It’s exciting, isn’t it?” says Doug Dome, who has been studying data’s impact on marketing for 30 years. Dome, who works as a marketing consultant and adjunct professor at University of Chicago’s Graham School, grows excited as he talks about the possibility of AI: the time it could save marketers, how it can bring companies closer to consumers and its potential to catch customers in stride, saving effort on the business and consumer side. As an integrated marketing communications professional, his excitement about the potential of AI has given way to the belief that AI will completely change branding, marketing, advertising and perhaps the world.
“Just think about all the innovations, all the promise of technology,” Dome says. “Is your life now really that much more convenient? Is it easier? I don’t know that it is. … I think in order to be able to fully benefit from a data-driven marketplace, marketers will have to take a broader perspective on problem resolution, and the tribal approaches that Pedro Domingos has articulated are the solution.”
Dome is referring to Pedro Domingos’ book, The Master Algorithm. This is the future of marketing, he believes. Dome animatedly spins his fingers around a circular chart within the book that explains the need to bring unique tribes—or philosophies—of machine learning together, each with their own algorithm.
So certain is Dome that AI is the future of marketing that he has banked his time and money on it with Core7, what he calls his “entrepreneurial sabbatical.” Core7 is, in theory, a marketing platform that applies AI to marketing ecosystems via a “master algorithm.” The algorithm would be licensed to brands and agencies, which he says would create a hyper-speed version of a fully integrated marketing ecosystem. However, Dome has been unable to get the company off the ground; investors have not yet been on board with the idea. It’s an ambitious goal, he admits, and when he first started pitching the idea two years ago, it was downright audacious. The Core7 team was developing the platform and algorithm and ready to go further, but thus far, Dome has been left to study AI from the outside.
Dome still believes he’s on the path to finding marketing AI’s master algorithm. “It may not be me, but it will be somebody like me that will ultimately develop an applicable master algorithm in marketing,” he says. “I’m disheartened to some degree, but at the same time I know I am on the cutting edge of where the marketing industry is headed. I know that philosophically, I’m there.”
What Is Marketing AI?
For many marketers, terms like AI, machine learning and master algorithm may seem akin to a foreign language. In Domingos’ words, the “master algorithm” would work much like a key that could open every lock. A professor of computer science at University of Washington, Domingos says this is the big difference between the machine learning he writes about—which functions as the limitless key—and traditional programming. To keep the comparison consistent, new keys must be created for every lock in traditional programming; if marketers want to track a certain subsegment of customers, they must create a new algorithm for each.
“The beauty of machine learning,” Domingos says, “is you don’t have to program the computer to do any of these things. The same algorithm will learn to do all of them depending on the data you give it.”
Domingos describes AI as a subset of computer science, in which computers can undertake reasoning and common sense tasks—such as vision and knowledge—which were formerly only undertaken by humans.
Stuart Russell, professor of computer science and Smith-Zadeh professor in engineering at University of California, Berkeley, describes AI a bit differently on his website: “It’s the study of methods for making computers behave intelligently. Roughly speaking, a computer is intelligent to the extent that it does the right thing rather than the wrong thing. The right thing is whatever action is most likely to achieve the goal, or, in more technical terms, the action that maximizes expected utility. AI includes tasks such as learning, reasoning, planning, perception, language understanding and robotics.”
Machine learning is a subset of AI that allows computers to learn the same way people do, only faster, without being explicitly programmed, Domingos says. Machines can rapidly change, grow and create when new data is inputted into the system. In theory, this means a program might be able to do years of work in the span of days or even moments. It is, Domingos says, the fulcrum of AI and what gives computers potential to learn, hold conversations, seem human and potentially create their own marketing algorithms.
“AI is the goal; AI is the planet we’re headed to,” says Domingos. “Machine learning is the rocket that’s going to get us there. And Big Data is the fuel.”
The central idea for Domingos’ “master algorithm” is to take algorithms from the five machine learning schools of thought (Bayesians, Evolutionaries, Connectionists, Symbolists and Analogizers) and meld them into one. The Core7 concept would shrink this down to an industry-specific basis, Dome says. For example, the automotive industry could have a single master algorithm, as the customer journey is essentially the same at each company. This master algorithm would, in theory, add efficiency, increase ROI and allow brands to develop a customized relationship at the consumer level that would revolutionize branding. While Dome’s dream has yet to be fulfilled, Domingos already sees an entryway within the marketing industry. He believes that in five to 10 years, machine learning will be used beyond marketing and across entire companies.
“The first can be segmentation … but then it spreads to everything else,” he says. “When you look at companies like Amazon and Google—the most advanced in machine learning—they use machine learning in every nook and cranny of what they do.”
In fact, Amazon has become so good at machine learning that a third of its business comes from a machine learning-powered function: recommended purchases. Similarly, Domingos says approximately three-fourths of movies watched on Netflix come from the company’s recommendation system, which also runs on machine learning.
“The recommended system is very famous at Amazon, but it’s one of many,” he says, calling this “quintessential machine learning.” “They’ve become good enough at this that they’re starting to roll out what they call ‘predictive delivery,’ in which they send you stuff before you even order it. They’re so confident you want it that they just put it on the truck. I’ve asked them, ‘What happens if I get this and I don’t want it?’ They say, ‘Well, we’ll just let you have it for free.’ This is how confident they’ve become in their ability.”
While Domingos says Amazon has yet to pinpoint exact future purchases, the company is adept at stocking items on the delivery truck with the knowledge that someone will order that item within hours.
This concept could solve a real challenge marketers have: hitting the customer “in stride,” not just having them come to you, but knowing when they stop and start, where they travel and what they need. Knowing their desires, more or less, and having the ability to communicate with them via AI chatbot programs or automated messages without wasting employee time. The potential of AI allows companies to use data already at their disposal to measure in real time, learn more about the customer and anticipate what happens next.
“Today is very much a race to who can develop the master algorithm first,” Dome says.
Marketing’s Quest for Singularity
“Our technology, our machines, are part of our humanity. We created them to extend ourselves, and that is what is unique about human beings.” - Ray Kurzweil, futurist, computer scientist and inventor
When Markus Giesler was a child, he was floored by the idea of the profoundly villainous HAL 9000, a conceptual AI from Stanley Kubrick’s “2001: A Space Odyssey.” He was so titillated by the idea that he and a friend tried to recreate a good-natured version of HAL in his own home. For weeks, Giesler would videotape his parents as they entered and exited rooms. He analyzed their language and noted their moods, realizing his AI would have to be tailored to his parents’ experiences to deliver the realism of HAL.
“About a month or two later, we had finally established a constellation that worked: every time our parents entered the room they were able to have a one-minute conversation with a computer. Not really the most elaborate chat but enough to impress them—and the occasional guests,” Giesler writes on his blog.
Giesler, who is the chair of the marketing department at York University’s Schulich School of Business and director of the Big Design Lab, researches AI concepts further down the path of his childhood creation, such as smart homes and driverless cars. However, humans were interested in AI long before his adventures with HAL, all the way back to antiquity before the Middle Ages, he says. There has always been a longing for what he calls “technology with a spirit.”
“It’s surprising to me that we’re only now beginning to see AI as a marketing construct and as something to look into from a marketing and customer experience design standpoint,” he says. “It makes sense for it to become more mainstream now when you consider the influx of AI algorithms, apps and mechanisms coming into everyday consumption, but artificial intelligence is not necessarily a new thing.”
What has changed is the awareness of AI, particularly in marketing. This awareness seemed to begin with a bang in 2012 with the infamous story of Target accidentally figuring out a young woman was pregnant before her father did by automatically analyzing her shopping habits and sending her advertisements for baby necessities. Now, perhaps startled by the technology’s abilities, companies have convinced themselves of AI’s impact. In a June 2016 report, Weber Shandwick found that 68% of CMOs report their company is “planning for business in the AI era” with 55% of CMOs expecting AI to have a “greater impact on marketing and communications than social media ever had.” This change in awareness may go a long way toward marketing and other industries accepting AI. Giesler says a shift in the decision-making process takes as much change in humans as it does in technology.
“I am most fascinated with AI in marketing when it’s invisible,” he says. “It’s one thing to talk about AI as this [creation of] applications that totally immerse consumers into these extraordinary experiences. It’s another to see how AI has invisibly crept into some of the most taken-for-granted aspects of everyday consumption to shape who we are as individuals, who we are as families, how we think about safety, togetherness and all this. One level on which we see that is cellphones having become an extension of who we are.
“AI is dramatically reshaping and redefining not only the market and what companies can or cannot do with customer experience, but who we are as individuals and groups,” Giesler says.
In a towering office building off of the Chicago River sits a notable example of AI’s current and potential capabilities. Narrative Science, a natural language generator, has become well-known by marketers for its ability to produce written stories within seconds based off analytics. The company’s AI can use data from Google Analytics, for example, and write sentences like: “New users spent 16 fewer seconds on your site than returning users did last month. This could indicate that your new users didn’t ﬁnd the information they needed or came to the site expecting something else.”
Katy De Leon, vice president of marketing at Narrative Science, says she couldn’t believe the company’s claim when she first read the job ad four years ago. “It just sounded incredulous to me,” she says. “I needed to talk to someone about it because I just couldn’t believe it.”
After four years of seeing AI in action, De Leon is a believer in not just Narrative Science, but in the potential for AI in marketing. AI has come at the right time with the explosion of Big Data, she says, and her company’s capabilities are especially mind-boggling at first glance for those on the outside. Narrative Science, born at Northwestern University as a collaboration between a computer science class and a journalism class, received coverage early in its existence when journalists at The New York Times and other publications were awed by a tool that could put together sentences from raw data—in this case, reports from sporting events. Now, the most lucrative customers of Narrative Science are in the government and the financial industry—think Fortune 1,000—as well as web analysts and small to medium-sized marketers.
Eyes across the industry are on the marketing tech landscape, which even De Leon admits is getting crowded and noisy. However, with increasing access to data, it’s never been more important for organizations to make sense of the noise. AI is another tool marketers can have at their disposal with potential for saving money, increasing efficiency and improving business.
“When you have 20,000 customers and you want to communicate with them as if you know them very well and … communicate something relative to them—something they care about—we can enable them to get to that level of personalization at a scale that wouldn’t be possible with people,” she says.
Where is Marketing AI Going?
Marketers should expect quick changes with AI’s potential to build upon and grow itself, experts say. Businesses and marketing departments are already vigorously moving ahead with the adoption of AI technology, according to Meabh Quoirin, co-owner and CEO of the Foresight Factory & Future Foundation. They are eager to see its promised benefits come to fruition.
“It’s not just about automation for automation’s sake, but if we can go faster, there’s more money to be made,” she says of the average company’s mentality.
How humans view technology, especially in marketing, has progressed over the past five years, Quoirin says, and it’s likely to keep progressing at breakneck speed. There are many possibilities for AI in marketing, health, entertainment and business; the technology is just starting to bear fruit, she says.
One possibility sure to entice across industries is what Quoirin calls “beyond human” AI, which can be used to “cheat death,” as well as add human bio-enhancements, prosthetics or implants. This could work well in the medical field, of course, but she says it may also work from a customer experience perspective. Marketers could find interest in tools for performance improvements for the average person; ways to burn calories, eat well, work faster and move better, especially considering the success of gadgets such as the FitBit.
“Broadly speaking, we tend to find that as soon as people are using [technology] like this in a context where it helps them get things done faster, they adjust to that convenience very quickly,” she says. “What we see is that it is a question of ‘when’ rather than ‘if’ with AI. But it will happen bit by bit. A lot of the things we worry about will just gently recede as we get used to being better humans.”
AI advancements may also change the concept of who we are and how marketers interact with humans and their technological extensions. Giesler says how consumers represent themselves online, how machines become an extension of who we are and whether marketers should market to these technologies once they gain a certain kind of sentience are all concepts he actively studies. “That’s wicked, right?” he says with a laugh.
Gieseler has done his own research on where humans end and where machines begin, which he says is an unbelievably fascinating and terribly scary new frontier to study. This inevitably leads to questions about how people live, how their habits are measured and how they’re watched by government-run AI technology, such as facial-recognition software—another budding AI concept. This brings to light existential fears of society becoming a bit too similar to George Orwell’s novel 1984, causing many to demur at the thought of AI’s rapid progress. Through all of these possibilities and theories, Giesler believes marketers can take center stage in redefining and renegotiating the boundaries of where the human ends and where technology begins. It’s an onerous duty filled with opportunity.
“We are the ones who best understand the human technological interfaces and how to design markets that are truly better than the sum of their parts when it comes to these redesigned interfaces,” he says.
Apple is the best example of this thus far, Giesler says. He’s assisted in Apple’s research and says Apple TV—a recent advancement that Steve Jobs called simply a “hobby”—is reshaping in-home entertainment and branding with AI concepts.
“For a long time, Apple adopted this top-box approach where you have the Apple TV box next to other cable boxes. That probably didn’t work,” Giesler says. “The difference came when Apple recognized that consumption is really more a matrix than an individual box with a person looking at a TV screen. If you want to conquer the living room, you really need to spread all through the home.”
Lead your marketing organization’s planning process to maximize success! Join Professor Greg Marshall of Rollins College in the latest installment of the AMA’s live training series. Get details and register today!
By seeing the market as more of a matrix, he says Apple cleared the path for marketers to use interfaces that let consumers better navigate their lives. Enabled by AI, Apple and researchers made these advancements after looking through the lens of today’s technology-enabled world.
“AI leads to changes in the way we do marketing, not just in the tech space literally, but also metaphorically, in terms of how we understand brands, customers and market segments,” he says.
Will Marketing Jobs Be Safe?
Upon hearing about AI’s capabilities, many will ask, “What’s the catch?” There are the existential fears expressed by Bill Gates and Elon Musk that computers could become too smart and take over the world. There are fears that AI could occupy the citizenry’s space too heavily and be seen as an invasion of privacy. Then there are palpable fears of AI taking jobs away from marketing and many other industries. After all, robots and computers don’t make a yearly salary.
According to a June 2016 report from Forrester, AI, machine learning, robots and automation will mean a net loss of 7% of U.S. jobs by 2025. The technology will mainly eat away at office and administrative support staff. New jobs, such as content curators, data scientists and robot monitoring professionals, will be created, but the losses will be greater than the gains.
“I think there will be an impact on jobs; we call this trend de-pop in the sense that working at large is going to change,” Quoirin says. “There will be competition for jobs. Equally, the new jobs will create new demands … We do see a shift in that.”
Even with fears of job loss looming in marketing and across other industries, Domingos says humans will still be necessary due to a paucity of data scientists, or those who automate the work of computer scientists and create AI algorithms. These algorithms have potential to take jobs—a factor of 1 million, when you talk about automating the jobs of computer scientists, Domingos says—from many people, but there’s a lack of data scientist talent.
“The war for talent is really raging,” he says. “One reason the demand has exploded and the supply changed quickly is you need people with a Ph.D.; that takes five years. ... The irony is a lot of the professors are moving to the industry level, which is good in the short term but it’s actually eating the seed corn. There are not enough people to train the students.”
This may come as a breath of relief for marketers, but Quoirin says marketers should expect the necessity for a transition of skill set and talent management to more creative and conceptual endeavors, areas where humans thrive over machines.
“Let’s not be too vanilla on this: If we take a sector like finance or retail baking, there will be an eye on how many tellers can be replaced,” she says. “Those numbers of cost cutting will have been done already. But let’s face it, without even a hint of what’s to come in AI, those jobs were under threat. Simple computer processing and mobile banking have already threatened those kinds of things. Artificial intelligence is much beyond that level of cost cutting. People are mostly thinking about how they can rechannel mundane jobs.” Although Quoirin believes AI will be “unstoppable,” she says humans will still be needed to interpret AI’s signals and numbers.
AI is expected to keep growing. Neuroscientist and author Sam Harris, who presented a TED Talk on humanity’s potential to lose control of AI, said on his podcast “Waking Up” that AI’s growth will keep advancing unless something much worse happens to society first. For this reason, Quoirin believes menial jobs will eventually be replaced by the robots, which may mean an alternate solution, such as a minimum salary for all, needs to be considered.
“There is, of course, also the future where we just work less,” she says. “And we get longer weekends. Wouldn’t that be fantastic?”
The Way Forward: Excitement or Fear?
“Some people worry that artificial intelligence will make us feel inferior, but then, anybody in his right mind should have an inferiority complex every time he looks at a flower.” – Alan Kay, computer scientist
AI’s marketing moment may be coming soon, if it isn’t already here. Domingos says Silicon Valley had its AI tipping point five years ago but kept it to themselves as a “secret sauce,” of sorts, for competitive advantage. Now, the proverbial cat is out of the bag and he says CEOs of Fortune 500 companies demand AI. “I don’t know what it is yet, but I know we need it,” Domingos says, doing his best CEO impression. However, he believes adopting AI will be easier for digitally native companies—such as Facebook and Google—or industries like marketing or finance where data has been essential from the start.
“The companies furthest along also happen to be in sectors where they have profit margins large enough for them for afford machine learning efforts,” he says. “If you’re Google and you [essentially] print money, you can afford to spend money on machine learning and you do. If your profit margins [are 1% or 2%], then it’s harder. They can only afford to do it so much because they don’t have money to do more.”
For the Doug Domes of the world, this makes AI look that much more enticing. Dome believes AI has “limitless” potential for profitability and says the positives of the technology will be immense, even if there are some ethical and moral bugs to work out.
In Giesler’s view, the negative predictions of AI have always been around; he’s always heard that his beloved HAL 9000 would be created in real life. However, despite advancement, he thinks AI is still far away from the ability to snuff out humans.
“There is something about being human that is unique,” he says. “There are simple mechanisms we can use to unmask the technology as what it is: a stupid series of algorithms that doesn’t really get it. That’s still pretty much the reality of everything we have around us.
“All the beautiful things we associate with marketing, they are and will continue to be the human actors and the human participants, not so much the technology,” Giesler says. “The beauty of real technology is that it’s like a mirror: We look inside it and what we see is who we are as human beings. Markets are human. The technology helps us get closer to the beauty of that principle.”