branding

15/12/2017

Will Robots Be the Next Graphic Designers?

We are proud to have teamed up with BRAND PACKAGING to share our 25-year’s design experience with retailers across North America.

BRAND PACKAGING’s editorial covers brand strategy, development, packaging innovations and trends and are the only publication in the US entirely focused on CPG branding through packaging. They take material/technology trends and consumer, design and retail insights, balance them with consumer research and brand marketing goals to create and deliver the information brand owners need to develop engaging packaging. The audience is typically influential marketing, design and executive decision makers who work at the earliest stages of new packaging.

Written by our US Director, Steen Tjarks, we are excited to share our opinions, experiences and knowledge to help the industry develop in 2018.


The next 40 years will behold a revolution to the way that humans work. Already in relatively early stages, advances in artificial intelligence (AI) have been staggering and concerning for the job market. By way of example, a recent study used IBM Watson to analyze 1,000 cancer diagnoses. In 99 percent of the cases, Watson was able to recommend treatment plans that matched actual suggestions from doctors. Furthermore, because it read and digested thousands of documents in minutes, Watson found treatment options that human doctors missed in 30 percent of the cases. With such results both exciting and alarming, many believe that for prescriptive, analytical jobs AI’s will rule supreme.

When considering the extent that AI’s may overtake creative industries’ the debate becomes far more complex. In fact, it even sways toward a symbiosis between man and machine which will give creatives of the future a competitive edge. Such an assumption comes from an analysis of our own creative process for packaging, in which we see human graphic designers continuing to play a key role.

In our long tenure as a design agency for ALDI, our work requires a constant balance between the bold and stereotypical. Packaging of this nature needs to be unique enough to be ownable but yet trigger feelings of familiarity with brands that consumers are already familiar with. In contrast, our recent work in the liquor sector requires purely original design that speaks to the brand’s emotions and story in a unique and innovative way.  With both of these types of design in mind we consider how an AI could disrupt or improve the standard design approach.


The Brief

Understanding a client brief is more than just the words used, but rather is a deep understanding of the client, their consumers and the psychology of the approver. Even the simplest briefs can be complex to discern, with much being left to interpretation. Often, we get briefs that are as simple as: we want something fresh and original, that will keep existing consumers but attract new ones’. In a scenario such as this how well would an AI fair?

First, AI would have to understand what fresh and original meant from the perspective of the client. True, an AI could easily evaluate the definitions of each word and cross reference them with both new and old designs the client is about to launch, but could it truly understand what the client wants? By definition this would require understanding the client’s tastes and their personality and while it is possible to gain inferences on these traits from social media, medical records, memberships and purchases are these truly representative of a person’s tastes? Research to date suggests not, especially since many X and Y gens do not use online channels as their primary mode of purchasing. However, with the viability of bricks and mortar stores ever decreasing, and the use of online media increasing, predicting a client’s personal and professional tastes in the future will become ever easier.

The verdict: Humans win now, AIs will win in the future.


Market Research

A core component in any good design process is a good market audit. This research stage requires the assimilation and evaluation of similar products on the market which are specific to the project on hand. Considering the direct competition’s brands is of particular importance so as to avoid similarities. In a best-case scenario, a good market audit can take half a day, depending on the availability of images online. If product images aren’t available online, in store audits are common practice. The lack of an online store, as is the case for ALDI, renders efficient and accurate online audits difficult, especially for an AI.  With most retailers offering a fully representative online product portfolio this limitation applies to but a few retailers, a case even less likely in the future.

Using AI for market research has already begun, despite the large operating costs. For instance, AMAZON’s Alexa (not considered an AI in its own regard) is already able to recommend consumer’s products and alternatives in a matter of seconds. Furthermore, the use of machine learning in predicting marketing and advertising trends is already gaining popularity, with processing times being in minutes rather than hours. Results from Intel’s Saffron AI support this, finding unparalelled savings in efficiency and unprecedented advances in predicting consumer tastes.

Verdict: Humans win for now (due to costs and availability), AI will definitely win in the future.


Design & Deliver

Artists and designers have always used tools to create their work: for Van Gogh it was the paint brush, for Rob Jannof, the designer of the Apple logo, pen and paper. Despite much of today’s design work being executed on machines, we have and always will appreciate an object carved or touched by the human hand. Furthermore, creativity is an ability in which one combines ideas, both stereotypical and original, to create a piece of art. Since AI’s are algorithm based they will be perfect in determining the former. Using the internet, social media and sales data AI’s will undoubtedly have the ability to quantify design trends specific to target markets. Most impressively though, they will be able to assimilate all of this data in mere minutes providing a serious completive edge over a human designer, and, but how would it ever fair at creating something original?

Recent testing has found that AI can certainly generate things that look like design, although whether these could be considered original pieces is a harder question. In particular, original works use expression and emotion to create uniqueness, both factors that AI’s can neither execute nor comprehend. Say, for instance, the client asks for a design that is on trend and uses bold colors, this is what an AI will deliver. What it will fail to do is consider and potentially include a new innovative piece of design in monotone that is crazy and different. In our experience, providing design options like this gives the client a wider set of options to choose from, many a time a benefit to all parties. Showing original innovative ideas is always good, it shows you have created something from scratch, it shows you have gone above and beyond and it is often praised by the client.

With it being unlikely that AI will have the ability to create design that invokes emotion in the next 50 years, it may be possible to view the machine not just a tool, but a partner or collaborator, with its own ability to create. Future designers are therefore likely to going to need to modify their roles from being designers to curators, using the machine-driven content to inform their emotive design work. Since design is equivocal to an art form, it is debateable if an AI will ever surpass a good designer’s ability to express a brand from the perspectives of real consumers.

Verdict: Humans will always win, although will greatly need the help of an AI to deliver and remain competitive.


With the design industry being highly competitive and relatively low paid it seems reasonable to question the feasibility of AI in the industry at all. Recent estimates from Cisco’s AI, Kim, suggest costs upwards of $50,000 USD per year for the most basic tailored system. If using an AI for more extensive design purposes a figure of over $250,000 USD seems more reasonable. With such high figures in mind, large design companies may need to re-evaluate their pricing structures to remain profitable. The extent that the cost vs. benefit of AI’s to such agencies is highly speculative, although efficiency and productivity will dramatically increase, thus equating to greater margins should an effective human-machine design process be integrated. In contrast, smaller agencies or freelance designers will be left unable to compete on efficiency, speed and productivity, leaving them in some regards at a disadvantage. However, what these designers will lack in efficiency they will make up for in commitment and love. As already stated having the human touch in design work, at a fraction of the price of the big competitors, will always be appealing to a potential client.


The End Verdict

AI will be fundamental in almost every area of our lives in the future. It will offer convenience, productivity and results but not true originality or emotive inspiring work. While there will undoubtedly be huge changes to the way design agencies operate, AI’s will never be able to replace or compete with creative, innovative designers. This will be especially so if both the agencies and designers adopt an approach that complements an AI’s strengths.

Noodles Signature