Are you panicked yet that artificial intelligence (AI) will soon put you out of work? Could a robot take your marketing job? Some of the brightest minds in Silicon Valley are warning of massive job displacement across the economy in the next decade.
But there remain good reasons not to be terribly alarmed. At least, not for a while.
First, the bad news: according to ThinkGrowth.org, “Between 9 percent and 47 percent of jobs are in danger of being made irrelevant due to technological change (in the next 15 years), with the worst threats falling among the less educated.”
Some panelists at SXSW this spring were even more apocalyptic. Bill Gates said, “AI is the biggest threat to the human race. I can’t believe more people are not worried about this.” Steve Wozniak added, “Fast machines will eventually get rid of slow humans.”
There’s no question the nature of work will continue to change, of course. Automation has been gradually displacing human labor since before the industrial revolution. And AI will expand the range of tasks machines can perform, through “smart” automation.
Yet the future for workers may not be so bleak after all, particularly in skilled trades and in creative professions like marketing. Here’s why.
Robots have been used in manufacturing since 1959. And it’s true, automation in general, and robots in particular, have had a significant impact on factory employment. The number of U.S. workers employed in manufacturing fell 39 percent from its peak in 1977 to 2012. Five million factory jobs have disappeared since 2000, partly due to trade but primarily due to automation.
However, those trends don’t quite tell the whole story. The U.S. has actually added one million manufacturing jobs since employment in the sector bottomed out in 2010. And growth is continuing. According to the latest figures from the Bureau of Labor Statistics (BLS):
“In February (2017), employment in manufacturing rose by 28,000. The manufacturing diffusion index increased from 50.0 in January to 65.4, its highest level since November 2014. A value above 50 indicates that more component industries gained jobs than lost them.”
Factories are having trouble finding workers—at least finding those with the right skills. In Minnesota, for example, “nearly 5,000 manufacturing jobs are unfilled — a number that will likely grow as more and more employees move into retirement.” And nationwide, Bloombergprojects, “Over the next decade, 3.4 million manufacturing jobs are expected to become available as baby boomers retire and economic growth spurs work opportunities… but a skills gap could result in 2 million of those jobs staying unfilled.”
How is it possible that employment may grow and factories may face (human) worker shortages even as robotics and AI technologies advance? Simple: Automation increases productivity (which increases societal wealth) and makes the U.S. more competitive globally. We’ll need more workers and more robots.
A recent U.S. government report—Artificial Intelligence, Automation, and Economy—predicts driverless automated vehicle (AV) technology may eliminate 2.2 to 3.1 million existing U.S. jobs. But any such job losses that occur won’t happen immediately or abruptly. They will be spread out over time.
Further, the report concedes that certain types of drivers (e.g., long-haul truckers transporting goods) are more likely to be replaced than others (school bus drivers transporting children, for example). The study also notes, “New jobs will also likely be created, both in existing occupations—cheaper transportation costs will lower prices and increase demand for goods and all the related occupations such as service and fulfillment—and in new occupations not currently foreseeable.”
And those projected job losses assume AV technology will become reliable and trusted. Though great progress has been made (driverless vehicles are being tested in several citiesbeyond San Francisco, Detroit, and Pittsburgh), some of the hardest work remains. As the expression among software developers goes, the first 90 percent of a project takes 90 percent of the time; and the last 10 percent of the project takes the other 90 percent.
AV technology will need to work nearly flawlessly before adoption becomes widespread. Business Insider has reported that lawyers are “salivating” over self-driving cars because they are “going to get a whole host of new defendants,” with deep pockets, in the event of any crashes.
Development of AV technology that works dependably regardless of weather, daylight, and other conditions remains challenging. As Gary Marcus, a best-selling author, entrepreneur, and professor of psychology at NYU, pointed out in TechCrunch regarding AI, “look for example at a driverless car, that’s a form of intelligence, modest intelligence, the average 16-year-old can do it as long as they’re sober, with a couple of months of training. Yet Google has worked on it for seven years and their car still can only drive — as far as I can tell, since they don’t publish the data — like on sunny days, without too much traffic.”
Still, robots and AI already have displaced some workers and will continue to expand into new jobs, particularly those that deal with things rather than people. It will likely be a long time before robots are trusted to care for children, or adults with special needs, but they’ve already been running warehouses for years.
Public policy will need to address those job losses, for example with displacement assistance and retraining programs. But standing in the way of AI and robotic progress would be counterproductive (literally); by increasing productivity, they raise living standards across society. Schemes like a robot tax are a bad idea.
Technology has eliminated wide swaths of employment in the past, from telephone operators and electric typewriter repairers, to photo technicians and video rental store cashiers. It’s now threatening various types of clerks, professional drivers, even insurance underwritersand appraisers.
But AI is more likely to change how marketers work than to replace them. It will supplement the efforts of human workers rather than take their jobs. Why?
First, consider one type technology already in wide use: marketing automation software. Despite the label, these applications don’t “automate” marketing; they merely enable marketing professionals to set up sequences of email messages which are then automatically sent out using (human) defined sequences and branches.
There are marketing professionals, agencies, and consultants who specialize in optimizing the use of marketing automation systems. In the words of Marketing Week, marketing automation platforms “don’t destroy jobs, they just change what jobs are needed.”
Second, there are several distinctly human characteristics essential to marketing that will likely prove vexing to reduce to mimic with silicon.
Interpretation: An AI-based tool like PaveAI can evaluate 16 million possible correlations within Google Analytics then produce a report showing the most significant findings. But it still requires a human to interpret the results.
For example, knowing that the highest conversion rate correlates with visitors who land on your home page on a weekday during business hours is about as unsurprising as any data point could be to a B2B marketer. But discovering the lowest conversion rate associated with a particular section of your website visitors often reach through organic search is far more interesting, and actionable.
Sentiment analysis presents another type of problem. Words like bomb, sick, mad, bad, and beast are generally considered negative terms to associate with your brand; yet all have, within recent memory, had a positive connotation in slang. People get that (hopefully). Machines will likely struggle.
Creativity: Marketing is an almost uniquely left brain and right brain profession. Data analysis, where AI can help, is of course vital.
But emotion plays a significant role in every considered purchase process, impacting both consumer and B2B buying decisions.
The creative side of marketing appeals to our emotions, and that side requires distinctly human creativity. It’s difficult to imagine, for example, even the most sophisticated AI systems coming up with something like E*TRADE’s invest in vests commercial.
Originality: AI can help marketers optimize current channels, but it won’t develop radically new ideas. For example, AI can help optimize and personalize email content—but AI never would have come up with the idea of using email for marketing in the first place (that was Gary Thuerk of Digital Equipment Corporation).
AI may help with optimizing messaging and timing on social networks. But it couldn’t have spontaneously computed Oreo’s famous dunk-in-the-dark tweet… Or suggested creating a profile for KFC’s famous founder on LinkedIn. And it certainly wouldn’t have invented a sporting event to support brand content marketing, as Red Bull has done with Crashed Ice.
Perspective: Not every question, in any realm of life, has a clear-cut answer. Even when looking at the same underlying data, reasonable and intelligent people can disagree, based on their beliefs, assumptions, experiences, and definitions—in short, based on their perspective.
For example, is it possible to accurately measure the ROI of social media marketing efforts? AI could provide an answer—and with the right data sources, even perform the calculations—but it couldn’t provide the perspective on the answer that a human thought leader provides.
In marketing content, it’s often the perspective that’s as interesting as the answer. It’s difficult to imagine an AI system weaving a narrative from a unique or interesting perspective. It’s even harder to imagine AI writing this post.
Persuasiveness: Great marketing in any form—text, visual, video—combines logic with emotion to move buyers to act. AI has logic literally at its core, but trying to teach AI to understand human emotions has so far been an enormous challenge.
An analysis by The Economist on the impact of robots and AI on employment suggests not only that the fear of massive job losses is likely overblown, but that in some cases automation may actually increase the number of jobs for humans. A study of the American job market from 1982 to 2012 found that:
“Employment grew significantly faster in occupations (for example, graphic design) that made more use of computers, as automation sped up one aspect of a job, enabling workers to do the other parts better. The net effect was that more computer-intensive jobs within an industry displaced less computer-intensive ones. Computers thus reallocate rather than displace jobs, requiring workers to learn new skills. This is true of a wide range of occupations…
“So far, the same seems to be true of fields where AI is being deployed. For example, the introduction of software capable of analyzing large volumes of legal documents might have been expected to reduce the number of legal clerks and paralegals, who act as human search engines during the ‘discovery’ phase of a case; in fact, automation has reduced the cost of discovery and increased demand for it. Judges are more willing to allow discovery now, because it’s cheaper and easier… The number of legal clerks in America increased by 1.1% a year between 2000 and 2013.”
The analysis also reiterates that almost every new wave of technology in the past has raised the specter of mass unemployment, only to end up creating more jobs than were destroyed. The term “technological unemployment” sounds like a concept Gates or Wozniak may have devised. The phrase was in fact coined by economist John Maynard Keynes in the 1930s. The total U.S. labor force more than doubled in the following five decades.
In marketing, AI will take over routine and data analysis-intensive tasks, but also create new opportunities for human employees—for example, in training and teaching AI systems. AI is already being used in areas like personalizing product recommendations and more granularly targeting advertising.
But AI requires human training, testing, and teaching both during the implementation phase and on an ongoing basis. Both human testing and human judgment are needed up front in terms of preparing AI platforms for the real world and determining when they are ready to go live.
A Harvard Business Review article points out the level at which AI systems are “good enough” varies widely by application; a mistake by Alexa or Siri in understanding speech and ordering the wrong item is annoying. A mistake by a self-driving vehicle may be fatal.
Once live, AI platforms—just like a human graduating from college and entering the workforce—need continued training over time to increase their capabilities and stay current with changing tastes and technology. And that means people, as explained in VentureBeat: “AI’s advancement up the value chain is only possible with the aid of human intelligence.”
Historically, technological advancements have always ended up creating more jobs than they destroyed. Today may prove to be different, but for now, it appears robots are more likely to be workplace assistants rather than job terminators. As a marketer, you probably don’t have to worry about robots or AI taking your job. But you will need to be prepared to work with these technologies to do your job better.