This column originally ran in Ad Age. I shortened it considerably before submitting it, and they are much better editors than me, so you can find the polished version there. My rougher, more meandering version is below.
A chief creative officer who I met with several years ago lamented about how much work was farmed out to her agency group’s centralized production house, and that within a few years, she wouldn’t be needed there. “Your job is the safest,” I told her, as the one thing the clients wanted was the head of creative pitching them big ideas. Everyone else had more cause for alarm, as clients weren’t as concerned with who did the work.
Now, I worry more for her, and our industry, and even our construct of humanity as we confront two forces: the increasing capabilities of artificial intelligence, and the feelings of disaffectedness that have been exposed brutally by President Trump. Artificial intelligence breakthroughs may be what lead liberal, American city dwellers to empathize with the president’s most steadfast supporters.
There are two types of artificial intelligence. Strong AI is typically what movies conjure – the likes of the voice in the movie “Her” that can constantly learn more and accomplish more tasks. Narrow AI, by contrast, is designed to accomplish a very specific task. My favorite example of narrow AI is Clippy, the maligned Microsoft Office cartoon paper clip; it theoretically could predict your needs and proactively offer solutions to certain tasks within select programs (even if it didn’t usually work quite that well).
Narrow AI is the immediate threat and opportunity. By harnessing AI-powered technologies to accomplish specific tasks, some marketers and their organizations will find new efficiencies and revenue opportunities. Much of this will be based off existing marketing automation software that becomes more effective by learning from every interaction, and from every new data point it ingests.
These narrow use cases will proliferate, and they will affect more areas of the organization. Consider human resources, and recruiting in particular. Software using AI will be able to learn from common traits held by top performers at an organization and more intelligently recommend new job candidates. The same AI that could potentially surface proverbial diamonds in the rough could also threaten to make a firm more homogenous by recommending people who resemble current staff. Even as such AI improves, human recruiters will be in control of vetting all job candidates. Such recruiters will have to be vigilant to reap the benefits of such software without getting swept up in the drawbacks.
AI will also threaten creative process. There are already numerous examples of AI creating works of art, from classical music that sounds more like Bach than Bach (as designed by David Cope) to original images of paintings that are indistinguishable from human-made works (powered in part by Facebook’s AI Research Lab).
Reflecting back on the conversation with the creative chief, she will be needed as much as ever, but the bigger threat to her team will come less from holding company outsourcing than it will from algorithms rapidly creating work. These infinite compositions, with guardrails in place so that they adhere to the brand’s guidelines and brief, will then be funneled to a different narrow AI assistant to test the performance of creative work through small media buys, so the marketer will nearly instantly identify the most effective ads for each customer segment.
As AI seeps into the creative process, I have two concerns. One is that people won’t know what is created by humans and what is created by algorithms. This already happens, as people have no idea when they’re looking at a forgery hanging in a museum – because the museum’s curators don’t know either. AI-powered software will be able to scale infinitely though so that the percentage of creative work created by humans steadily diminishes in relation to the share of work created by machines.
The greater concern is that people won’t care. Consider an exhibit of real works of Van Gogh mixed with computer-generated interpretations. Will people spend any more time looking at the real ones? Will they demand to see only the Van Gogh originals? Or will they just want to be inspired by art, regardless of who or what created it? Toothpaste ads are hardly “Starry Night,” but how will that make those of us in advertising feel if people just don’t care about the role we human beings play in the process? If people care less about the human provenance of the epitome of creative arts, it’s hard to imagine consumers concerned with who designed a 10-second digital video ad.
For the creators of this work, including those involved in the process from selling it to signing off on it, we have to be prepared for the emotional impact that these changes will bring. We will risk becoming disaffected and disconnected. We can’t afford to say that what has happened in the rust belt’s manufacturing plants and Appalachia’s coal mines won’t happen to us. Many of us will run the risk of becoming disaffected and disconnected. To make matters worse, David Brooks will write columns about us. It really can get that bad.
In the process, some good might emerge. We will be humbled. We will learn empathy in the hardest way possible. We will look for ways to avoid turning disaffection into bitterness and hate. We will have to use all of our creative talents to do the one thing that ad execs were born and bred to do: ensure that what we do matters.