Can AI be More “Creative” than Humans?: Launch of AI Advertising Copywriter
Can AI even substitute an advertising copywriter – a “creative” profession?
Advertising and AI
Uses of AI technology have been spread across various industries. Marketing, which is all about people’s psychology, is not an exception. Advertising agencies compete in developing their own marketing “tools” using AI, to effectively analyze target consumer behavior and to efficiently plan optimal media budget allocation.
In developing advertisements, there are three major functions:
Strategic Planning (“What to Say”), Media Planning (“Where to Say”), and Creative Development (“How to Say”).
AI technology had been good “friends” with Strategic Planning and Media Planning, where it is important to analyze past consumers behaviors statistically and achieve efficient and effective planning. However, it was not as easy when it comes to Creative Development. Creative Development had been thought to be the final “raison d’être” of advertising agencies; However, that is not the case these days. How should advertising agencies deal with AI technology? Is “creative” not a “ unique expertise” of humans anymore?
Dentsu launched AICO, the AI copywriter.
Headquartered in Tokyo, Dentsu Inc., the world’s largest advertising agency as an individual company[2], is apparently trying to catch up with this megatrend, instead of escaping from the reality. In 2017, Dentsu Inc. launched the AI copywriter, AICO, which can generate a huge amount of advertising copies at a single command[3].
VIDEO http://www.dentsu.com/business/showcase/ai_planners.html
It is important to note that they are not trying to “replace” the traditional human copywriters; Instead, they are trying to “co-work” with AI copywriter and to become more efficient in developing advertising copies. Developing ad copies are not an easy job. Copywriters think day and night, and come up with tens of, sometimes, hundreds of options, most of which will be discarded in the process of internal reviews or client reviews till they get the one and only option which is finally going out to the public. Considering the copywriting is the product by an advertising agency, this process of “product development” takes weeks, or even months, depending on the capability of both clients and agencies to collaboratively create something new. Having said that, from clients’ perspective, it will be better if they can get the same quality of advertising copies in a faster mode under the supervisions from the real human copywriters. Here, AI copywriter AICO serves as an innovative tool for process improvement.
What will happen in a long run when AI technology improves further in terms of not only the quantity but also the quality of the output? Do clients not need advertising copywriters anymore?
The only answer I can get from what Dentsu is doing is that the complicated task of “integration” will be left to humans. Dentsu organized AI MIRAI (“mirai” means “future” in Japanese), a cross-functional project team who collect all the resources inside/outside the company related to AI and promote their use of AI into various marketing opportunities[4]. AI technology can be used directly by clients themselves, but to integrate various opportunities created by AI and package them into IMC (Integrated Marketing Communication) which is specific to a client’s problem is something AI alone won’t be able to do.
Recommendations
As an MBA candidate who worked in the advertising industry for several years, I would like to suggest a few ideas to better tackle the rapidly changing situations around AI.
For a short term, they can expand AICO into other languages and own the space of a traditional advertising agency who is familiar with AI. AI, coming from its nature, is especially compatible with digital marketing space, and there are already implementations of AI copywriting in digital advertising, such as Google[5] and Alibaba[6]. However, one uniqueness of AICO is that it has machine-learned thousands of copywriting works of Dentsu’s copywriters. If an AI copywriter from a traditional advertising agency can differentiate itself from Google’s or Alibaba’s, then the key differentiator will be the quality of the inputs.
For a long term, end of the day, what they should do remains the same : Retain and nurture “human talents” for ad creative. As I analyzed above, in order for the advertising agencies to be kept needed by clients, it is important to notice that the more the AI technology improves, the better the quality of AI copywriting will be, the more important the human’s skill of integration will be. Moreover, the key differentiator of copywriting quality would be the input of base copywriting data created by Dentsu’s copywriters. Hence, it is crucial to keep nurturing creativity as well as integration skills using AI technology.
Yet, they cannot completely rest assured. What if AI start to learn how to “integrate” marketing communication into one packaged solution? What if clients start to “internalize” creative expertise within their company using AI and stop outsourcing to agency?
(Maybe it is a good idea to ask AICO to come up with an appealing concluding sentence.)
[799 Words]
References
- “15 Business Applications for Artificial Intelligence And Machine Learning”, Forbes Technology Council, 2018, https://www.forbes.com/sites/forbestechcouncil/2018/09/27/15-business-applications-for-artificial-intelligence-and-machine-learning/#180cdf4a579f, accessed November 12, 2018
- Annual Report 2013, Dentsu Inc., 2013, http://www.dentsu.com/ir/data/annual/2013/bds/index.html?fbclid=IwAR3B_bijQ0GU8Lz0dYQJHMC9FbSjMdy6qtIge682FTjBr6pHvdKtNpy85m0, accessed November 12, 2018
- “AI Planners MAI & AICO”, Dentsu Inc., http://www.dentsu.com/business/showcase/ai_planners.html, accessed November 12, 2018
- “AI MIRAI – A Dentsu Project Team”, Dentsu Inc., http://www.dentsu.com/news/ideas/ai_mirai.html, accessed November 12, 2018
- Peter, Adams, “Google brings AI-powered ad suggestions to AdWords campaigns”, Marketingdive, April 18, 2018, https://www.marketingdive.com/news/google-brings-ai-powered-ad-suggestions-to-adwords-campaigns/521549/, accessed November 12, 2018
- Lisa, Lacy, “Alibaba Says Its AI Copywriting Tool Passed the Turing Test”, Adweek, July2, 2018, https://www.adweek.com/digital/alibaba-says-its-ai-copywriting-tool-passed-the-turing-test/, accessed November 12, 2018
First off: what a great last sentence! I’m sure that AICO could come up plenty of options, but maybe not the best option (only humans can do that…for now!). The questions posed at the end of this post are through provoking in the context of the increasing role that AI will play not only in advertising but also in consumer decision making. I would offer that the greatest opportunity to accelerate adoption of AI in advertising would be from increasing penetration of AI in consumer decision making. Were this to occur, we could see a situation where automated ads adapt quickly to the preferences of automated consumers, hastening the pace of AI’s entrance into advertising.
Thank you for your article!
You raised many interesting points on the applications of AI within the advertising agency. While I was initially skeptical, you have convinced me that AI can be a valuable technology even within a ‘creative’ field like advertising. Going forward, I particularly agree with your point on the importance of the data that is used to power the AI technology. In order for the algorithms to develop content that will resonate with customers, data scientists will have to appropriately design which data is used as a input.
Great post Sada! I think this delves into the challenge facing many industries right now, where do the humans go when computers take over? It is a scary thought for many displaced workers if even in what is considered the most creative job, advertising, workers are at threat of losing their jobs. One question, though, is what are the inputs AI is using to generate its creative ideas? I imagine that their outputs are still very limited within the box of their inputs, whereas many ad agencies thrive off thinking outside the box (many times with ads that could seem completely unrelated to its industry). I’d also love to know more about the timeline of AI in ad agencies. Do you think this is a threat in the next 50 years or is creative development coming more rapidly?
Thank you for the extremely interesting article on AI in a totally new context – marketing! The questions you have raised around integration of AI copywriting with human copywriters is at the heart and core of the subject, to distinguish what a machine can do vs. humans. I am curious to learn more about how AI actually fosters more creativity? Particularly, if AI is driven by historical copywriting trends, how does it create a completely new type of copywriting? What are the inputs used for it to test out different copywriting methods? Moreover, how will advertising companies differentiate themselves if a machine is doing the task of copywriting – as you have pointed out, perhaps the human copywriters and their art to distill that information will continue to be the key differentiating factor. This research and article, almost reminds me of the Gap case study, where we try to understand the role of a creative director and if a computer and machine learning can truly replace the “art” part of the creative director, while helping the “science” part of their job.
Great insight Sada! Though slightly different, your post reminds me of our GAP case where we explored the role of “big data” replacing a traditional creative director in product development. Although I believe there is a role that AI and big data play in informing marketing insights, I do not think AI can ever replace the human touch in creative development. Marketing works its magic by creating human connection. It’s not enough to know how people behave – you have to be able to anticipate how they want to behave, and most importantly how they feel. I don’t know how lines of code could replicate that connection when offering creative recommendations. Perhaps AI can help in the execution, using testing to inform which variations of creative resonate best, but I firmly believe it can never replace the initial creative process. Hopefully this means you have job security!
Super nice article, thanks for sharing! I loved your breakout of the digital marketing stack and agree that AI has started to affect many of its levels. I unfortunately also agree that creativity won’t be safe for long… Convolutional neural networks are already trained on millions of examples of a particular type of data (say, synthesizer sounds, or… advertising copy) and learn parameters that are particularly useful in that class of data. They then can generate examples on their own by initializing with random weights and optimizing to create output that look like the training data. Here’s a few examples of different creative coding initiatives 🙂 : https://experiments.withgoogle.com/collection/ai
Sada, thank you for your post on AI in Marketing! What’s curious to me is that Creative Development (“How to Say”) in advertising is not just the copy in isolation but also the context. To the point of other commenters, AICO and AI in copywriting will be heavily influenced by the data they’re fed, but part of what makes advertising interesting is how the same words can send a different message depending on the context, imagery, and cultural references of an ad. Going back to your great comment for the Nike case, the symbolism and emotional provocations of an ad are things that humans can more easily connect and create, and so the copy must be married to the context. Where do you see AICO and AI going with this? Could that be where humans continue to be relevant for some time?
I struggle to understand how AI in creative development would actually work. For example, would AI be capable of proposing that we use a visual with people that are homogeneous vs. diverse without actually being provided those as options? I am curious if we have attained this technology.
I think management and selection of creative assets produced by humans is much more manageable task efficiently done by AI. Presuming appropriate training of the algorithm, we can attain consistency in decision making and hopefully a reduction in personal bias. For example, predicting which of the given visuals will do better in a certain geography seems well within the realm of what AI can help us do.
Lastly, can we count on AI to make those drastic changes such as those that we have seen in the Dove campaign or is there something uniquely human about the same?
Thank you for the post, Sada! It is clear that AI will have an increasing impact on the advertising industries in the years to come. However, I do not believe that the advances of technology would enable the companies to be fully independent from advertising agencies. I would say so because if companies started to use only AI for advertising, the demand for more edgy and transformative ideas (that can only be produced by humans) would increase, creating necessity for ad companies, which can innovate!
Thank you for sharing, Sada! While I agree that advertising agencies should be concerned about the improving capabilities of AI in this space, I still believe humans hold a couple advantages, at least for now. Machine learning requires extensive data that inevitably contains information on the past. Since one aspect of a great advertisement is being currently relevant, do humans have an advantage in terms of speed of data consumption? For example, if an ad were to feature a reference to a current pop culture phenomenon, how quickly could enough data incorporating the relevant content be fed to an algorithm for it to use this data effectively? My guess is that a human would beat an algorithm to the punch in creating an effective ad with reference to this piece of pop culture, though I may be wrong. Another aspect I thought about was humor and emotional content in general. I would love to know how AI-produced ads compare to those produced organically in terms of producing cleverly comedic or emotional content.
As a former creative copywriter, I would love it if AI stays as far aways from advertising as possible. While this might seem great on paper, digital interfaces are still struggling to accurately translate simple sentences from one language to another, let alone write a whole creative copy. We are so far away and honestly I am not sure if I want us to go there. Advertising is all about genuine and deep understanding of consumers’ culture, mindset, and behavior. Good copywriters are those capable of immersing themselves in the inspiring and magical complexities of the human mind. I am very skeptical if we will ever reach a day when machines can do this with the same level of creativity. And if this happens, then as a client I would just buy me one of these machines and say my goodbyes to advertising agencies.