Students were tasked with illustrating a Hindi idiom, “आ बैल मुझे मार” (“Come, bull, hit me,” meaning to invite trouble) through two separate visuals: one literal and one figurative. Afterwards, they described their own illustrations via detailed prompts to generative AI (ChatGPT-4) to replicate the same images. The AI delivered aesthetically pleasing results but consistently missed the subtle interaction, like a child pushing a glass from below while a woman drank, which humans naturally depict using causal and spatial logic. This exercise highlights how AI lacks true comprehension of coordinated and intentional actions despite advances in image generation. However, Gen- AI tools will learn from their mistakes and our prompts, potentially reshaping the creative process of independently generating solutions versus asking for them.

Visualising idioms means turning sayings into pictures that show both their literal and implied meanings. The challenge is to make the idea clear so children understand them through pictures, while also capturing the idiom’s essence with creativity. It is about blending clarity, culture, and imagination into a scene you can see.

In this experiment, students illustrated the Hindi idiom “आ बैल मुझे मार” by first creating two images: one for denotation (literal meaning) and one for connotation (implied meaning). In the second phase, they reverse-engineered both images using text prompts to generate a similar image with generative AI. The goal was to see how well AI could match a human-made illustration showing complex, simultaneous actions.

Humans naturally understand intention, causality, and body coordination, sequencing actions logically to tell a story. AI (Chatgpt 4), however, works by assembling elements from patterns in its training data without real comprehension. This leads to missing subtle but crucial relationships, revealing its difficulty in accurately portraying interdependent actions that require spatial awareness and physical logic.

Create a square cartoon-style illustration of an Indian mother and a kid, in two scenes within the same square frame, divided by a gentle wavy horizontal line.

Scene 1 (top-left quarter): Mother (short, stout, round features—round face, round eyes, smooth nose, round bun, bulky arms—wearing a plain green gown with round neck and elbow-length sleeves) is drinking water from a steel glass, glass in her left hand, upper rim touching the top of her nose, head tilted back, eyes closed. A lanky kid (protruding eyes, small nose, curly hair with a bunny-tail ponytail, loose plain red T-shirt, loose plain blue shorts) stands to her right, left arm up, finger pressing the bottom of the glass, water spilling over her face. Kid has a mischievous look—wide eyes, tongue out at corner of mouth. Only their top halves visible.

Scene 2 (bottom half): Mother chases the kid. Her left hand has just thrown the steel glass at him, which is in mid-air toward the running kid. Her right hand clutches her gown near her thigh, raising its bottom edge slightly above her ankle. Her expression is furious—furrowed eyebrows, wide protruding eyes, mouth wide open shouting. The kid is mid-hop, running away terrified—wide eyes, mouth open—slightly overlapping into the top-right quadrant.

Composition notes: The divider line is a gentle wave from middle left to three-quarters right. Background white. Minimal line strokes. Exaggerated cartoon style, not childish. Single darker patch for shadows. Force lines show motion of glass, hopping, and chase.

AI tools, like the scissors in the image, may seem helpful, giving quick and neat results. To novices, they might even appear fast and clean at first glance. No wonder they are slowly improving and getting closer to what we desire. But in the long run, they should not replace our ability to think independently and critically. When we as designers surrender our thinking to tools, we risk becoming quick solution seekers rather than problem solvers, cutting off the very branch we sit on.

As artificial intelligence-generated content (AIGC) (Lou, 2023) become increasingly integrated into design workflows, their role in shaping communication design pedagogy must be examined. This research explores how students’ reliance on the internet influences their creative process, from idea generation to execution, and how AIGC tools interpret prompts to generate visuals.

In a design education assignment, master’s students in communication design illustrated an idiom’s literal (denotative) and figurative (connotative) meanings. Initially, students were instructed not to use AIGC tools for image creation. However, they searched the internet and ChatGPT for verbal explanations of idioms, often encountering visually described examples that influenced their work, sometimes directly shaping their final illustrations. Consequently, some student solutions became variations of the references they found.

Findings suggest that students view external references whether from traditional searches or AIGC tools as integral to their design process. Rather than perceiving reliance on such sources as a limitation towards independent thinking, they regard it as a source of inspiration.

In the second phase, students described their illustrations in prompts to AIGC tools, which generated visual representations for comparison. None of the AIGC outputs fully met students' expectations, failing to deliver the intended representation, though students acknowledged their superior rendering quality. Results varied based on students’ ability to craft effective prompts and the AIGC tools used. Engaging with AIGC at a later stage enabled students to critically assess its role as an execution aid rather than a solution provider.

As AIGC tools advance, their influence on design education may deepen, raising concerns about increased dependence, cognitive offloading, and shifts in conceptual thinking, potentially reshaping the creative process itself.

Lou, Yongqi. 2023. ‘Human Creativity in the AIGC Era’. She Ji: The Journal of Design, Economics, and Innovation 9 (4): 541–52. https://doi.org/10.1016/j.sheji.2024.02.002.

Artificial Intelligence-Generated Content (AIGC), Design Education, Communication Design, Visualizing idioms

“Sar mundate hi ole padna” (सिर मुँडवाते ही ओले पड़ना) is a Hindi proverb meaning to encounter obstacles, misfortune, or failure immediately upon starting a new task or venture. It translates to facing, “misfortune at the very outset” or “setbacks at the beginning”.