Combined Intelligence: Your New Superpower

It seems lately that headlines about artificial intelligence volley between two competing narratives. The first narrative would have you believe that artificial intelligence is bad. From accusations published in The New York Times and The Wall Street Journal that YouTube’s algorithm favors extremism, to software using biased data to predict future criminals employed by sentencing judges, to Amazon’s failed attempt at artificial intelligence assistance for recruiters, the need to thoroughly examine how machines are influencing our decisions has been pushed to the forefront of our attention. On the other side of the debate are those who believe that artificial intelligence is capable of amazing things. After all, it is already helping to find and diagnose diseases, like heart failure, depression, or pancreatic cancer.

It’s an important debate mostly because it forces attention on using artificial intelligence wisely. Policy makers and industry leaders are and should be working hard to ensure that these rising technologies are implemented fairly and ethically at a global level. Increasingly, the media industry is also examining the systems it has created, the behaviors it incentivizes, and the impact, however unintended, it has on individuals and society. However, the endless consternation over artificial intelligence as a force for good or evil ignores the most basic practicality – artificial intelligence is not human. And when it comes to connecting with other humans – an essential job of advertising – that matters quite a bit.

Hayden Lewis, one of The Richards Group’s digital strategists and author of last year’s examination of the state of AI, defines artificial intelligence as simply “the ongoing and evolving process of how humans train machines to think like humans.” I like this description because it centers on the role of humans in the process. Machines can only do what is asked of them with the data available to them. They do not have intent. The humans who train them do. They are acting as a mirror for our assumptions, biases, and native structures.

This is why it’s critical to understand how these systems are built and to understand when, why, and how they fail at their intended task. We have to understand where artificial intelligence stops and when human intelligence needs to take over. Our new superpowers lie in how brands pair their best tech with their best humans. Leveraging the strengths and weaknesses of artificial intelligence will yield the strongest partnerships.

 

Efficiency and Optimization

The first major impact of artificial intelligence is efficiency. In our industry, artificial intelligence has been helping to automate routine tasks that are often time-consuming and repetitive – in particular, helping marketers efficiently find, target, and cater to their unique audience.

Chart showing U.S. marketers’ goals when using AI

There are already studies that show using artificial intelligence in this way can improve campaign performance – both on efficiency and effectiveness.

Chart showing the effectiveness of human vs. machine-driven campaigns

Source: AXIOS

As targeting more specific audiences gets easier, machines make it more efficient to optimize and personalize creative, further improving effectiveness. Here at The Richards Group, we’ve used dynamic creative optimization to increase new donor acquisition rates for The Salvation Army giving campaigns. In 2017, this increased revenue for their fourth quarter digital advertising campaign by about $600,000. Beyond driving the business, leveraging dynamic creative optimization resulted in several other benefits for the agency, such as an increased ability to test new creative elements with fewer barriers to production and less creative wear-out because of the increase in versions produced. Realizing these benefits is often a result of shifting time, not replacing people – allowing the algorithms to do much of the routine work, while the humans focus on objectives, strategies, testing, and thoughtful asset development.

As machines take on more routine tasks, they free up human resources for more complex thinking and problem-solving. This will only be amplified as machines are able to take on more complex tasks. Gartner, a global research and advisory company, suggests that “by 2022, one in five workers engaged in mostly nonroutine tasks will rely on AI to do a job.” Human strategy and creativity will become increasingly vital to getting the most out of this technological transformation. In fact, the World Economic Forum believes that by next year, creativity will be one of the most important skills in our workforce. The next wave of artificial intelligence’s influence will be on how we explore, discover, and create, in particular reinventing how strategists glean insights and how agencies create.

Comparison of top 10 AI skills in 2020 and 2015

Source: World Economic Forum

 

Discovery

Marketers today are often swimming in data without time to make sense or use of what they have. Humans, because of time constraints, have to guess at what data to use, with no guarantee that the exploration will be fruitful. Machines can solve this problem by looking at all the data, identifying areas of interests for further exploration. While humans will still need to be the ones to find the insight, machines will broaden the universe from which we gather them.

New data sets are another exciting opportunity provided by the age of artificial intelligence. Computers can comb through not just statistical data, but behavioral data, unstructured image or text data, and qualitative data to enable us to learn more. For example, aggregation and analysis of customer reviews could accelerate iterations of product or experience design. Image recognition could better understand how brands appear in customers’ lives. Machine learning could detect new patterns in customer interests that make them better targets for our ads.

While machines will make many new insights possible, it’s also important to understand their limitations. They are built by humans to make particular decisions based on existing data. This predisposition toward what’s known is visible in today’s algorithm-based experiences. The Pinterest feed “knows” my tastes so well that it no longer helps me discover new things. Brands keep showing customers the last product clicked because it’s still the most reliable predictor of a sale, even when it’s not an ideal customer experience.

It also has consequences. Photographers no longer have complete control over how they tell their story, especially if they hoped to do so chronologically. YouTube influencers change their formats to get more time spent and money, not because it’s right for their brand. This environment drives brands toward similar aesthetics, personalities, and creative ideas. It’s counter to differentiation. As brands or influencers try to introduce new messages and ideas, they will have to find ways to break out of the recommended path set forth by algorithms and artificial intelligence. In short, the machines need humans to take the next step.

An abundance of data and well-honed principles of success also makes doing something new and untested increasingly scary in our metrics-driven world. It’s particularly hard when data sets and models are built from data sets with relatively low variability. They are simply less useful in telling us the impact of a bigger risk. To expand the usefulness of our models, testing plans need to include tests of new approaches to increase the range of data available in models. This is similar to how humans make mistakes or explore extremes as part of the creative process and can help us recalibrate our decision-making models. Maintaining a testing mentality will remain an important trait of successful brands.

Another essential role for humans is to provide the happy serendipity of unexpected connections that fuel discovery. Computers have difficulty applying learning to new environments or connecting seemingly unrelated data sets. Strategists of the future will need a strong handle on how these tools can and cannot be applied to effectively direct machine systems toward new explorations. Hypothesis generation and explaining why will still largely be human endeavors.

 

Creation

The human and robot team will also reshape creative development. There are already interesting ways that artificial intelligence is helping in creative processes. For example, Chef Watson and Bon Appétit created an app that helps home chefs invent recipes. The essential human and artificial intelligence work together to birth new recipes. The app arguably failed at regularly giving clear instructions, but its structure and outlandish suggestions still inspired novel thinking. The value of the app was not that it always produced delicious recipes, but that it could push our culinary explorations to unexpected places.

Photo of cookbook
Source: Amazon.com

At The Washington Post, Heliograf, an artificial intelligence reporter, increases the newspaper’s coverage, particularly of local, statistics-driven news like elections or sports. It assists reporters by writing articles or identifying potential stories early, reducing errors, and freeing up time for in-depth research.

Google has created an artificial intelligence, Benjamin, that can write screenplays. Its first attempt was the sci-fi film Sunspring. It’s now moving on to directing with its new project Zone Out.

In each of these examples, artificial intelligence provides the basic elements of a successful creative product, but humans are required to curate and create something that other humans enjoy. This is an important limitation in the near future. As evidenced by Benjamin’s rather clichéd sci-fi foray, artificial intelligence may simply give us the rules to break in order to stand out – a useful first round of brainstorming. It could solidify which direction a creative team wants to head, and allow them to quickly test ideas or visualize a style. For example, Airbnb uses artificial intelligence to rapidly prototype web design, and tools like Deep Dream Generator or Prisma allow artists to easily reinterpret an image in various artistic styles.

Image of woman
Source: Prisma Photo Editor

Artificial intelligence is also allowing us to bring our brand points of view to life in increasingly human ways with chat bots, digital voice assistants, and increasingly customized experiences. How we use these tools says something about our brand.

Beyond the way these experience look or sound, there is an even more interesting opportunity to design artificial intelligence systems with a brand’s purpose and perspective in mind. What is its goal? How does it make mistakes? Does it have a sense of humor? When should it not be used? As voice skills, chat bots, and personalized subscription services continue to proliferate, creating a brand experience that customers will choose over others will in large part be decided in those key moments.

One important opportunity for creativity and differentiation is in how we hand off customer interactions between machines and humans. ESPN recently partnered with IBM to provide fantasy football players with recommendations based on both structured (player statistics) and unstructured (news, analyst reports, and sentiment) data. Interestingly, this partnership didn’t result in a tool that automatically optimized a player’s lineup, but rather enabled them to make smarter decisions while still giving them final control over their teams and success – a critical part of the fantasy football experience. Success in employing artificial intelligence in customer experiences will depend in part on how empathetic the experience can be, a theory Google is applying with a dedicated team in its Empathy Lab. For this reason, The Richards Group can assist clients with everything from how a brand should sound and act via artificial intelligence to developing the intelligence and experiences themselves.

Partnering with artificial intelligence will change the way strategists and creative teams work, often freeing their time to work on more interesting projects. Working well with artificial intelligence will still require astute human judgment. There will be times when principles of success based on huge amounts of data should be applied to briefs, creative, or media plans. But as Stan Richards says, sometimes you should set aside the brief. The same philosophy is worth considering in how brands and agencies approach the actions and rules set forth by artificial intelligence. The machine is only as smart and insightful as the humans trained it to be. Human connection is a vital part of what we do. So while we can and should learn what we can from our machine partners, we should not assume they are always right. Advertising is still very much both a science and an art, now fueled by a new kind of partnership.

How to harness the superpowers of combined intelligence:

  • Know how and why your artificial intelligence partner was built. Use its strengths, but know its blind spots.
  • Free humans to do what humans do best, understanding and connecting with other humans.
  • Use machines to cover more ground, both with data and creative approaches.
  • Keep an open mind and a testing mentality.
  • Design artificial intelligence with empathy.

The post Combined Intelligence: Your New Superpower appeared first on The Richards Group.

Bookmark the permalink.

Leave a Reply