By Bryan Samuelsen
Indulge me in a thought experiment. The year is 2050, and every action you take is observed, analyzed, and rated. These ratings contribute to your overall “social score.” If your social score is low enough, you are deemed “untrustworthy.” Your air travel is restricted. Public venues display your personal information so that others know you are untrustworthy. Your children may be prohibited from attending certain schools. You will be labeled “untrustworthy” for the next three to five years, or until you make enough public restitution to improve your score.
If you find this concept terrifying, you will be even more terrified to know that this program is currently being piloted in China.[i] How is it possible? This proficiency in social analysis is a result of artificial intelligence, popularly known as AI, which can analyze and applying huge amounts of facial-recognition data.
Luckily, most uses of AI are not so intrusive. However, China’s ominous employment of AI raises thorny questions about the technology’s capabilities. In particular, the advertising industry is currently wrestling with AI-related debates. Perhaps most gripping is this question: How smart will AI get? When it comes to advertising, will the technology be relegated to handling logistics, or will AI eventually become so intelligent that it can actually create its own original advertisements?
By understanding AI’s current capabilities and its entrance into the advertising sphere, we can better anticipate the future of AI and how to prepare for it.
What Can AI Do Right Now?
To understand AI’s current capabilities, we first need a working definition of AI. Most Americans do not fully comprehend what AI is or what it does. As demonstrated in Figure 1, only about a third of Americans classify Facebook photo tagging, Google Translate, and Netflix recommendations as AI.[ii] In reality, all of those systems rely heavily on AI to function. In contrast, social robots—the classic example of AI—were twice as likely to be classified as such.
Figure 1. Statistics from Center for the Governance of AI, Future of Humanity Institute, University of Oxford, Jan. 2019.
Relatively unknown definitions of AI versus ANI contribute to the above American classifications. Almost all current AI can be characterized as “artificial narrow intelligence” (ANI), which can perform only certain tasks and is incapable of reasoning through problems it was not designed to solve.[iii] Although these systems are not the kind of AI that you see in science fiction, they are still technically AI. So never fear; I, Robot is not an accurate depiction of today’s AI. The higher, self-aware forms of AI in such movies have not been invented yet—if they ever will be.
Even though the capabilities of today’s AI may be limited when compared to sci-fi conceptions, they are still astounding. Apart from automating media buy (bidding on ad space), which is now commonly relegated to AI, the technology can be used to write copy, create audio samples, and assist designers with video and image production.[iv]
In one of the most interesting cases, the Associated Press began in 2014 to transition some of its writing to AI, without any human intervention. For example, quarterly earnings reports—traditionally both an urgent and an unbearably tedious job—are now produced by AI. The same is true of Yahoo’s fantasy football reports.[v]
Can AI Design Creative Ads?
We have established that AI can perform some simple tasks autonomously. However, we must now move to a question fraught with implication: Can AI create its own ads without meaningful human input? This is a difficult question to answer. As of now, AI has not successfully designed an ad. However, in 2018, Lexus released the world’s very first AI-scripted commercial, which depicts a fatherly engineer who watches fearfully as the artificially intelligent Lexus that he built survives a harrowing crash test. However, there is an important caveat: the script was not a complete story and had to be interpreted creatively by humans.
In the first AI-written commercial, this engineer builds an artificially intelligent Lexus. Does AI writing about AI feel eerie to you? Source: Adweek.
To ensure that the commercial was a success, Lexus hired Oscar-winning director Kevin Macdonald. Ad agency The&Partnership made tweaks throughout the process, and though the script was indeed penned (figuratively) by AI, the AI in question was IBM’s Watson, arguably the world’s most advanced artificial intelligence. After being fed a slew of award-winning ads, Watson returned what was essentially a hodgepodge of data points linked into a tenuous story. These data points were then woven by Macdonald into a recognizable commercial.[vi]
Indeed, Adweek noted that the Lexus ad certainly pulled the right strings but that the story itself was not particularly cohesive, nor was it compelling. On the bright side, Watson’s work offered insights into what makes a successful ad.[vii]
Perhaps a commercial-length video was a little ambitious, but would AI fare better by creating a static ad? Kuaizi Technology in China uses AI to generate thousands of unique ad designs nearly instantaneously. This technology “can choose design templates, ad copy, and pictures for the advertisement,” with minimal input from the advertiser. Moreover, it then serves different ads to different target audiences and uses real-time data to adapt the ad based on what worked best previously.[viii]
In the Future, Will We Still Need Human Creatives?
So far, we have discussed how AI is quite common and is already capable of tasks like extensive facial recognition, copywriting, and mass-producing simple ads that dynamically change based on the user’s response. On the other hand, we have seen that AI thus far is limited to specific tasks, and even the most potent AI in the world struggled to create an ad that resonated with humans.
Will AI ever replace the human creative? Most opinions hold that the answer is no—that there will always be a need for a “human touch.” Here are three such opinions:
Dan Eckrote, managing director of Mindshare:
“We have to balance algorithmic machine learning with humanity, to still allow emotions and culture to influence the decisions we make. Technology is extremely effective in enabling scale, efficiency and effectiveness, but it’s critical to ensure there is still a human lens applied.”[ix]
Michael Tripp, Lexus general manager for Europe:
“None of this is designed to undermine the creative process but if we can add variables like 15 years’ worth of award-winning ads or luxury ads and data into the creative process then it liberates creatives to be more subjective. I’m very optimistic that it will complement and augment the creative process and not undermine or replace. We’re fully committed to man plus machine… Maybe we should call AI the creative teammate.”[x]
Reece Medway, media/entertainment specialist for IBM Watson:
“The magic of storytelling will always come to life in the human creative process. Using Watson to identify the common attributes for truly award-winning creative work is an example of how man and machine will collaborate in the AI era.”[xi]
These comments paint a rosy picture, but when “even a standard microprocessor available for $200 today runs at 10 million times the speed as a human neuron and computers can memorize more pieces of information in 1 second than a human could in a lifetime,”[xii] how confident can we really be that AI will never truly replace us? Is creativity in advertising beyond those capabilities somehow, or are we in denial that, one day, human intelligence of every stripe could be effortlessly replicated?
What Can We Do to Prepare for the Future?
Having engaged humanity’s collective existential fear, we now have to admit that we are relatively powerless at the present time. Too much about the future is unknown; who knows what breakthroughs in technology are waiting in the wings or what insurmountable barriers we might run up against.
However, what we can do is stay informed. Too many Americans and advertisers are illiterate in the language of AI, but we are seeing positive signs. The presidential campaign of Andrew Yang might have been short lived, but his platform of a universal basic income to help America transition into an automated future shows a growing willingness to address these issues.[xiii] In addition, the youngest generations are showing a technological literacy that portends changing times ahead.[xiv]
The topic of AI is certain to stay relevant for years to come as companies like Google and Microsoft experiment with its use in business processes.[xv] Some excellent sources to help you keep up on the latest include AI Trends,[xvi] which looks at AI in business and society; Science Daily,[xvii] which republishes relevant science news from various journals; and a variety of corporate and institutional blogs, such as OpenAI[xviii] or MIT News.[xix]
It is unlikely that we will have robot overlords anytime soon, but the reality is that artificial intelligence is already here, and it is only going to get more advanced. Advertising is not the only field it will disrupt. In fact, it will disrupt every field. Will we be ready for the disruption?
[i] Charlie Campbell, “How China Is Using Big Data to Create a Social Credit Score,” Time, August 14, 2019, https://time.com/collection/davos-2019/5502592/china-social-credit-score/.
See also Tara Francis Chan, “China’s Social Credit System Has Blocked People from Taking 11 Million Flights and 4 Million Train Trips,” Business Insider, May 21, 2018, https://www.businessinsider.com/china-social-credit-system-blocked-people-taking-flights-train-trips-2018-5?r=UK.
[ii] Baobao Zhang and Allan Defoe, “Artificial Intelligence: American Attitudes and Trends,” Center for the Governance of AI, Future of Humanity Institute, University of Oxford (January 2019). https://governanceai.github.io/US-Public-Opinion-Report-Jan-2019/executive-summary.html.
[iii] Andreas Kaplan and Michael Haenlein, “Siri, Siri, in My Hand: Who’s the Fairest in the Land? On the Interpretations, Illustrations, and Implications of Artificial Intelligence,” Business Horizon, November 6, 2018, https://www.sciencedirect.com/science/article/pii/S0007681318301393?via=ihub.
[iv] Rhoda Sell, “The Future of Advertising: Artificial Intelligence & Creativity,” Medium, July 2, 2018, https://becominghuman.ai/the-future-of-advertising-artificial-intelligence-creativity-522e969d194b.
[v] Ross Miller, “AP’s ‘Robot Journalists’ Are Writing Their Own Stories Now,” The Verge, January 29, 2015, https://www.theverge.com/2015/1/29/7939067/ap-journalism-automation-robots-financial-reporting.
[vi] Jennifer Faull, “Lexus Reveals Ad ‘Created by AI’. Is It a Gimmick? No. Will It Win Any Awards? Probably Not,” The Drum, November 16, 2018, https://www.thedrum.com/news/2018/11/16/lexus-reveals-ad-created-ai-it-gimmick-no-will-it-win-any-awards-probably-not.
[vii] David Griner, “An AI Tried to Write the Perfect Lexus Ad. Here’s a Scene-by-Scene Look at What It Was Thinking,” Adweek, November 19, 2018, https://www.adweek.com/programmatic/an-ai-tried-to-write-the-perfect-lexus-ad-heres-a-scene-by-scene-look-at-what-it-was-thinking/.
[viii] Gang Chen et al., “Understanding Programmatic Creative: The Role of AI,” Journal of Advertising 48, no. 4 (September 2019), https://doi.org/10.1080/00913367.2019.1654421.
[ix] Jon Lafayette, “Media Planning’s Next Big Change Agent: AI,” Broadcasting & Cable, August 13, 2018, https://www.broadcastingcable.com/news/media-plannings-next-big-change-agent-ai.
[x] Faull, “Lexus Reveals Ad ‘Created by AI’.”
[xi] Stephen Johnson, “Watch: The First AI-Scripted Commercial Is Here, and It’s Surprisingly Good,” Big Think, November 20, 2018, https://bigthink.com/technology-innovation/lexus-ai-commercial.
[xii] Kaplan and Haenlein, “Siri, Siri.”
[xiii] “The Freedom Dividend, Defined – Yang2020 – Andrew Yang for President.” Yang 2020, accessed February 24, 2020, https://www.yang2020.com/what-is-freedom-dividend-faq/.
[xiv] Lafayette, “Media Planning’s Next Big Change Agent.”
[xv] Alec M. Beresin, “Microsoft Is Challenging Google’s DeepMind With a Brand New AI Program,” Observer, July 13, 2017, https://observer.com/2017/07/microsoft-artificial-intelligence-google-deep-mind/.
[xvi] “The Business and Technology of Enterprise AI,” AI Trends, accessed February 25, 2020, https://www.aitrends.com/.
[xvii] “Artificial Intelligence News,” ScienceDaily. accessed February 25, 2020, https://www.sciencedaily.com/news/computers_math/artificial_intelligence/.
[xviii] “OpenAI,” OpenAI, accessed February 25, 2020, https://www.openai.com/.
[xix] “Topic: Artificial Intelligence,” MIT News. Massachusetts Institute of Technology, accessed February 25, 2020, http://news.mit.edu/topic/artificial-intelligence2.