Adapt or Die: How AI Will Redefine Creativity in Marketing

Created on iPhone6s with Prisma

What AI Will Mean for Creatives — and How You Should Prepare

Unless you’ve been living under a covfefe, you’ve probably heard that we’re living in the Age of Artificial Intelligence. Most creatives and marketers file AI in the hype bucket with good reason. It’s hard to imagine artificial intelligence equating to creative intelligence. But AI is coming for us all, and those that are unprepared to seize new opportunities will become irrelevant.

To the casual observer, the impact of AI might be limited to self-driving cars and gadgets that talk back to you. But researchers at Yale and Oxford recently surveyed 352 machine learning experts from 43 countries to get their view on how quickly AI will surpass human ability in a number of professions – including those that require creative thinking. Here’s what they came up with:

“Researchers predict AI will outperform humans in many activities in the next ten years, such as translating languages (by 2024), writing high-school essays (by 2026), driving a truck (by 2027), working in retail (by 2031), writing a bestselling book (by 2049), and working as a surgeon (by 2053).” Source: “When Will AI Exceed Human Performance?” by Katja Grace, John Salvatier, Allan Dafoe, Baobao Zhang, and Owain Evans

Unfortunately they didn’t ask the experts about “Creating Award-Winning Advertising Campaigns,” but it’s worth watching what is already in motion to get a handle on where things will go next.

What Are Some Examples of Creative AI Already in Play?

Anyone that takes an AI design application like LogoJoy for a spin will be underwhelmed. The AI is supposed to learn from your choices how to improve the options offered, but let’s be honest, the results aren’t very good — but they are fast.

LogoJoy sample designs

The same goes for initial attempts at using AI to write long form copy like news articles. Many major news outlets are trying out AI to write sports and financial stories, ostensibly to give “real” writers the time to work on more complex material. But, in 2016, Engadget gave Automated Insights’ Wordsmith a full training run and came to the conclusion that it couldn’t grasp enough context to be useful. Interestingly, the Engadget author saw a lot more potential in the short term for Wordsmith to write ad copy.

Persado, backed by Goldman Sachs and Bain Capital, has taken an early lead in the AI ad copy writing race. But Persado doesn’t “create” new copy — it uses data to generate alternatives for banner ads and email from lots of different angles and then tests them to see what works. So far they’ve notched customers like Hilton, Pottery Barn, and Verizon.

Another early player in the natural language generation space, Narrative Science, hasn’t yet focused on ad copy but they do create personalized customer communications for applications like financial statements. And startups like MarketMuse and Scoop.it are bringing AI to the content strategy space to create topically relevant messaging.

AI has played a role in video and animation for years, but is beginning to take on new tasks that will speed up production dramatically. Take for example Wibbitz, used by news outlets like USAToday to create video shorts automatically based on text story inputs. Microsoft’s research group is experimenting with AI that autocompletes animation originating from hand-drawn sketches.

Adobe, the 800 pound gorilla in creative software, has made huge investments in its Sensei AI but has only teased users with previews of what might be possible — though features in apps like Photoshop have been employing basic AI for years. Smaller upstarts like Prisma and Arista are already bringing AI to advanced rendering techniques for photography, video, and animation that you can use on a mobile device.

While AI is just getting started in many areas, it’s beginning to make major progress in web design. Although (mostly) autonomous website creation tools like TheGrid.io are achieving mediocre results, “guided” AI design platforms including WIX Artificial Design Intelligence, B12, and FireDrop have been successful by combining AI with varying levels of human intervention.

Marketing automation has long been more manual than automatic, but several companies are trying to change that with AI. Cortex, a specialist in social media, is using AI to automate engagement for clients like The Ritz Carlton and 1-800-Flowers. And Adgorithms is attempting to build a multi-channel marketing optimization platform with its Albert AI.

In the immediate term, some of the most effective uses of AI aren’t in creative output but in new approaches to creative decision making. With open source software like R Studio, even a novice data scientist can apply machine learning for predictive modeling, media mix modeling, and better customer personas with hierarchical clustering.

From Experiments to Practice: How Will AI Actually Impact Creativity?

AI won’t be stealing any creative jobs tomorrow, or even make your current job much easier. The software is still too experimental and is only effective in the hands of an experienced user. But in the next few years, things will begin to change. Here’s how:

1. AI will make you smarter.

Though marketing organizations and creative agencies all use data and analytics to some extent, most decisions today still come down to subjective hunches. AI will become critical to decision making and results tracking by delivering better understanding of our target audiences and where to deploy resources (e.g., more accurate media mix modeling).

2. AI will improve the creative process.

AI will engender new forms of automation to add speed and accountability in the creative process. In the short run, AI will accelerate production tasks like layout and versioning. In the longer term, AI will bring new efficiencies to project management, time tracking, and accounting. And AI will speed up digital development in big ways by bringing down barriers between designers and developers (see for example UIxard’s Pix2Code). This will cost some front end developer jobs, but create many more in backend engineering.

3. AI will improve your creative output.

AI offers the possibility of making bad creative pretty good and good creative even better. To start, more efficient processes will mean more time for creative thinking. But AI will have a direct impact on creative output too. Even if you find it hard to believe a computer could create a winning ad concept today, you should be prepared for some really useful advice very soon. With access to the complete history of advertising and lots of expert training, machine learning will eventually crack the code of what makes for good creative.

Recall the creative account manager’s favorite axiom: “Good, Fast, or Cheap: Pick Two.” For those ready to embrace the change, AI will enable better decisions, more efficient processes, and better creative output — and finally make good/fast/cheap a realistic possibility rather than a client’s dream.

When Will All This Happen?

It’s a matter of when, not if, AI can make useful contributions to all aspects of the creative process.

We’re currently in the first, experimental stage of AI’s impact on creative and marketing. Based on what’s happening already and the pace of AI development, we should see three phases overall:

  
    
    Timeframe
    Creative Quality
    Primary Benefit
  

    The Era of Experimentation
    Now
    Pretty bad
    Better decision making
  

    The Age of Augmentation
    5-10 years
    OK
    Production efficiency
  

    The Dawn of Autonomy
    10 years+
    Getting good
    Better creative work
  

These kind of predictions are necessarily oversimplified but useful nonetheless — even if just to get you prepared for what might be coming. The key variables to determine the rate — and effectiveness — of AI adoption are quality and volume of classification. That is, how quickly AI applications can absorb the kind of high quality and/or high volume feedback that is required for machine learning to be effective in context.

“Figure 1: Aggregate subjective probability of ‘high-level machine intelligence’ arrival by future years,” excerpted from “When Will AI Exceed Human Performance?”

(Many smart people think that creativity will be the final frontier for machine learning. For a summary of McKinsey’s exhaustive set of predictions, see 120 Machine Learning business ideas from the latest McKinsey report by Théo Szymkowiak and McKinsey’s own Tableau Public visualizations. See also Tony McCaffrey’s There Will Always Be Limits to How Creative a Computer Can Be, published at HBR.org.)

How Can You Prepare?

AI will empower those who are prepared to change.

When it comes to creativity and marketing, you should think of artificial intelligence as just a tool — not a sentient autonomous system that will steal your job. But that doesn’t mean that anyone is safe.

Within the next 10 years, creative and marketing organizations will begin to show signs of human-computer integration — and often it won’t be intentional or obvious. Traditional relationships, like that between Art Director and Copywriter, might alter dramatically or disappear altogether. Instead, AI will make it possible for creative roles to blend and put many more people into the Creative Director seat. The need for original ideas, brilliant copy, or amazing illustration won’t go away, but future creatives will need to think and act across disciplines.

In marketing organizations, anyone unfamiliar with advanced analytics and basic machine learning concepts should be concerned for their future job security. As in the agency environment, marketers will see a blending of responsibilities that will empower those who are prepared, but it will require a high level of proficiency in data science to survive — and thrive.

Appendix: An Incomplete Catalog of Creative AI Available and in Development

Copywriting

Persado

Narrative Science

Automated Insights Wordsmith

Content Strategy

News Writing