In 1986, thousands of typesetters took to the streets of London as the English capital’s leading publisher was set to replace 90% of them with a computer. While before, designers worked in tandem with analog typesetters and typographers—who assembled and turned newspapers’ copies, layouts, and designs into printed matter—new publishing software allowed them to feed their work directly into printers. Typesetters lost to automation in the end and were soon obsolete, paving the way for the modern graphic designer.
“A Robot Is After Your Job,” read a 1980 New York Times headline.
Decades later, those very graphic designers face a similar crisis as generative AI threatens their future. Over the last few months, tech firms have released tools that can auto-generate design copies, visuals, slide decks, wireframes, and “bring a creative vision to life” with a click of a button.
Punch in a few words about your brand on Canva, such as a “bike shop,” and its new AI tools will instantly produce sales decks, logos, and social media posts for you to use. Adobe, similarly, will soon roll out updates to its creative suite that will not only let anyone outsource their graphic design duties to an AI but also allow them to professionally tweak their photos, and videos without ever learning Photoshop. Wix Studio has launched AI text creators in its editor, allowing designers to generate website copy without leaving the platform.
Why so much concern around seemingly helpful tools? OpenAI, the research firm behind ChatGPT, conducted a study on the impact AI will have on jobs and found web, graphic, and design interface designers were potentially some of the most vulnerable professions. So yeah, the concern isn’t coming from nowhere.
How agencies are using AI now
Although Jessica Walsh, a designer and founder of the New York-based agency, &Walsh, believes creatives won’t have to worry about AI taking over their jobs anytime soon, she says their focus will shift more towards what today’s art directors do. The difference will be that “instead of directing humans, we’ll be directing AI to shape our vision,” and as AI becomes capable of automating designers’ workflows like end-to-end branding, more and more jobs “will be focused on AI prompt engineering and manipulating the output of AI.”
Walsh and her team are already experimenting with AI technologies in brand campaigns. On one recent project to rebrand a platform promoting nuclear energy as a tool against climate change, &Walsh used Dall-E, OpenAI’s text-to-image generator, to brainstorm ideas and generate thousands of images for the project’s backgrounds, case study themes, and typography. The agency’s designers then curated a few from this pool and further tweaked them to match their intentions. Agencies like Ogilvy Paris have used AI tools similarly.
Similarly, Jeff Turkelson, a strategy director at design firm Artefact, agrees AI taking over tedious tasks will augment designers at agencies, not replace them outright in the near term—and enable them to focus on higher-level decision-making and directing the design process. Turkelson, who used to spend hours online hunting down visuals and metaphors, now often turns to Microsoft’s Bing chatbot to quickly sift through the internet’s vast resources, and Notion AI, a writing assistant, to clean up his copy.
While designers are divided on how far-reaching this new wave of AI products will be, many agree that it will cut down the time to test and prototype a new idea.
With generative AI tools like Midjourney, another text-to-image generator, and Uizard, which lets designers instantly create UI mockups with AI, “we can quickly create multiple product iterations, compare the results, and return to the drawing board,” says Romina Kavcic, an Austria-based design lead. “This saves us weeks and even months of work.”
One Discord designer, for example, built a ChatGPT-powered Figma plugin that automatically generates the chats based on the number of people and messages and the topic, rather than manually populating his mockups with dummy conversations.
What the future of design might look like
In a research paper published in March 2020, researchers at the Harvard Business School argued that AI won’t undermine the principles of design thinking—which is both people-centered and abductive—but will overcome limitations in scale and scope, enabling designers to be more creative and personalize their work to an extreme level of granularity.
However, the paper added, AI will profoundly change the practice of design, and as designers automate their workflows, the human side of the design will increasingly become an activity of sense-making, and inch closer to management and leadership.
“As machines will do the tasks and propose solutions, designers will be left with the key question: ‘Does it make sense?,’ one of the paper’s authors, Roberto Verganti, a design theory professor at the Harvard Business School, told Shaping Design. “No machine can address this.”
Indeed, the only aspect of a designer’s job that AI hasn’t yet encroached on at all is the human one, from understanding the client’s needs to capturing the audience’s emotions. “Emotional empathy is a fundamental aspect of design,” says Tim Smith, a design director at NewTerritory, “perhaps even exclusive and unique to a human.”
But the industry will need to tread carefully
At the same time, the transition to AI software may not be as smooth as some hope. Researchers fear that the lack of engagement by graphic designers in training AI models, which is largely spearheaded by computer scientists, could overwhelm the field of design with overly functional, anodyne approaches. One June 2022 research paper published in the Design Research Society on graphic design and AI warned that this threatens to de-skill the profession and spawn a second tier of ‘non-professional’ designers, “particularly within less creative work that emphasizes fast turnover and functional artifact production.”
In such a scenario, Dr. Yaron Meron, the research’s author, and a design lecturer at the University of Sydney, expects professional design roles to adapt rather than go extinct, and possibly become more skillful in different ways, such as being better and more creative at prompting the most out of AI technologies.
Designers I spoke to agree if everyone drew inspiration and materials from the same set of models, it may lead to a rise in repetitive designs, but at the same time, they’re optimistic this would mean clients will come to expect more original thinking and unique content from studios, helping both further crystalize their role, and drive the intellectual skill forward—just like how Mac, Adobe, and Macromedia software did when it automated several manual processes in the computer revolution.
Besides, for now, AI-generated content has proved to be an ethical minefield for design agencies. Since these technologies are trained on existing human work without acquiring copyrights, there’s a chance brands that are using them may be indirectly infringing on an artist's or a designer’s rights. (Turkelson cites this as a major concern, and is hopeful that the steps taken by Adobe Firefly could be the path forward.)
Pau Garcia, the founder of a data-focused design and research studio, Domestic Data Streamers, has another concern: tech companies aren’t transparent with the material used to train these algorithms. That makes it harder for clients to feel confident that agencies created AI work with images that are fair use.
Ultimately, Dr. Meron adds, AI is just another tool and technological step for designers. Graphic design is first and foremost about communication—so a deep knowledge of the audience, brief, and context a designer is working within is far more important than the tools.
“If we look at what Neville Brody and David Carson were doing with [desktop publishing] software back in the 1980s, in what way was that different from what Milton Glaser or Paul Rand were doing a generation earlier? And how do they differ from those designers that produced all those marvelous 1920s art deco posters, or from even earlier, say, William Morris?” posits Dr. Meron. “I would argue not a lot.” The work isn’t different, even as the technology changes.