In a 2022 paper I argued that AI research into graphic design was built on a flawed premise – predominantly emerging from engineering, it treated graphic design as a set of mechanistic functional tasks, ignoring critical design aspects such as the brief, the audience, the communicative intent, and the creative process that precedes any artefact (Meron 2022). I also argued that graphic design’s own weak research discourse was partly responsible: if we don’t define what we do with sufficient rigour, others will define it for us (or ignore it).
Nearly four years on, most of it holds. The most sophisticated AI layout systems developed since 2022 still cannot accept anything resembling a design brief. They can generate spatially plausible layouts and reflow text across aspect ratios. But, what they can’t do is reason about why a particular layout conveys authority rather than playfulness, urgency rather than calm, or how a piece of communication should position itself relative to its audience and the competing motivations of diverse stakeholders. The evaluation metrics used to assess these systems measure geometry, not meaning. No benchmark asks whether a generated design actually communicates. (My subsequent collaborative research on AI in design education confirmed the same pattern from a different angle – even in pedagogical contexts, the tools reproduced this functional flaw (Meron & Tekmen Araci, 2023)).
The research discourse problem I identified in 2022 remains unresolved – but it’s compounded by a second issue. The practice itself is becoming increasingly opaque.
The industry picture is more complicated than either the doom narratives or the growth figures suggest. Industry sources put global graphic design services revenue at around $55 billion and growing at roughly 4% annually – figures that almost certainly vastly underestimate the discipline’s true economic footprint. However, while revenue is healthy, employment figures for graphic designers look flatter.
So what’s going on? The answer, I think, is that these kinds of figures seem to capture only the visible layer of ‘named’ professional graphic design services. But graphic design doesn’t behave like a bounded profession; the practice is much more widely distributed than that. For example, Canva (valued at over $65 billion) is clearly part of the same disciplinary and industrial landscape. But it’s even more complex than that.
Graphic design is a dispersed practice – operating across and within other industries, embedded in organisations that would never describe themselves as design businesses, with practitioners working under numerous diverse job titles, many barely resembling their origins or actual work. When it comes to government and industry classification, two practitioners with identical training, overlapping skills, and similar outputs can be classified in entirely different occupational categories depending on how their work is labelled. Compounding this, Australia’s occupational classification system was frozen from 2001 until its overhaul in December 2024, spanning the entire expansion of interactive digital design without meaningful structural revision (the UK and US systems suffer from equally confusing, but perhaps less acute issues). As job descriptions shift with industry fashions, practitioners disappear from the graphic design count entirely. The practice isn’t shrinking; it’s becoming invisible to the instruments that measure it.
This matters because what you can’t see, you can’t research – and what you can’t research, you can’t budget for, educate for, or adapt for. This matters all the more because, historically, graphic design has proved notably agile in responding to new technology. Even now, some Australian industry data suggests AI adoption in creative professions such as content creation, software development, and marketing runs at 22-25%, while graphic design sits at 10%. That’s a discipline whose relationship with technological disruption differs markedly from adjacent professions, and the reasons for the divergence remain almost entirely unexamined.
There are three reasons for this. Firstly, is graphic design’s own historical weak theoretical underpinning (HEFCE, 2014) and paucity of research (Corazzo et al., 2019). But this statistical invisibility mirrors a longer pattern of disciplinary marginalisation of the discipline, even by other design discourses, sometimes (like the engineering AI researchers), treating it as merely artefactual and decorative and even ‘too arty’ and unscientific (Brown et al., 2024). Finally, there is the emergence in the twenty first century of a series of grand theories of design, focused on procedural, sandboxed, risk-averse reformulations of the discipline; some – like design thinking – were aimed primarily at non-designers and within which disciplinary distinctions (including graphic design) had little space.
These grand narratives, or Neodesign (Hernández-Ramíirez & Meron, 2025), found purchase partly because graphic design had already been conceding disciplinary authority – being defined and perhaps even defining itself through its outputs rather than its process, its tools rather than its intelligence, its aesthetics rather than its communicative purpose.
There is a certain irony in the fact that businesses initially adopted design methods precisely because designers thought differently. Then the grand narratives began suppressing that difference, speaking the language of business, in pursuit of corporate legitimacy. One wonders if business has ended up with the critical and iterative versions of innovation that the consultants promised? Now AI reproduces the very functional, artefactual model of design that the profession itself helped to normalise.
Graphic design’s greatest vulnerability is not from technology. It is the continuing fragility of a research discourse that has not adequately articulated what the practice, in all its breadth and diversity of industry application, actually does. That case still needs making, because there is not one aspect of design, the creative industries, or even further afield, that isn’t impacted by graphic design – and at the moment, the practice is hiding in plain sight.
Brown, B., Buchanan, R., DiSalvo, C., Lee, K., Mazé, R. & Triggs, T. (2024) Introduction. Design Issues. 40 (3), 1-5. doi: 10.1162/desi_e_00763.
Corazzo, J., Harland, R. G., Honnor, A. & Rigley, S. (2019) The Challenges for Graphic Design in Establishing an Academic Research Culture: Lessons from the Research Excellence Framework 2014. The Design Journal. 23 (1), 7-29. doi: 10.1080/14606925.2019.1682446.
HEFCE (2014) Research Excellence Framework 2014: Overview report by Main Panel D and Sub-panels 27 to 36. London, Research Excellence Framework.
Hernández-Ramíirez, R. & Meron, Y. (2025) Neodesign: The Loss of Craft, Imagination and a Playful Attitude. In: DRS LearnXDesign 2025, 22-24 September, University of Aveiro, Portugal. doi: 10.21606/drslxd.2025.096
Meron, Y. (2022) Graphic design and artificial intelligence: Interdisciplinary challenges for designers in the search for research collaboration. DRS2022: Bilbao. doi: 10.21606/drs.2022.157
Meron, Y. & Tekmen Araci, Y. (2023) Artificial intelligence in design education: evaluating ChatGPT as a virtual colleague for post-graduate course development. Design Science. 9, e30. doi: h10.1017/dsj.2023.28.
Communication & Graphic Design Research 