Live2 Nov 2026Singapore artists draw lines as AI mimics work in seconds →
← Back to stories

Singapore artists draw lines as AI mimics work in seconds

Source: CNA Tech

Singapore illustrators and animators are tightening contracts, watching software updates and testing anti-scraping tools as generative AI turns prompts into finished-looking images.

Singapore artists draw lines as AI mimics work in seconds
SGAI Daily

A short prompt now yields a restaurant menu, an ad image, an event poster or a book jacket — work that once consumed hours, weeks or longer. Singapore's creative professionals have answered with fascination, irritation and sometimes revulsion, and the arguments keep returning to permission, copyright and the worth of handmade work.

Automation in studios is nothing new. Bernhard Schmitt, an associate professor at NTU's School of Art, Design and Media, notes that machine learning has aided animation for decades; the fresh shock is software trained on sprawling archives of other people's output. Graphic designer Pann Lim, founder of the agency Kinetic Singapore, sees reactions from flat refusal to eager uptake, plus a middle group using the tools to brainstorm.

The Woke Salaryman's co-founder Goh Wei Choon objects on procedural grounds: models ingest creative labour without asking. He also rejects the analogy that a machine learning from images resembles a person absorbing influences — human artists, he argues, sift what they meet through their own priorities and taste. Illustrator Farihath Samsul has watched her commissions reworked by others with AI; animator Jolene Tan, whose work has shown on MRT platform screens, has met generated sequences tracking existing pieces almost shot for shot, with only hair or clothing changed.

Defences so far are piecemeal. Samsul's terms bar clients from running her artwork through image generators. Comic publisher Difference Engine, co-founded by Felicia Low-Jimenez, tracks new AI functions in everyday software, refuses AI-produced submissions and pledges that contributors' files will never enter a model. Others rely on Glaze and Nightshade from University of Chicago researchers: the first distorts pixels against style mimicry, the second spoils unauthorised training. Neither undoes what has been learned, and Lim doubts such shields can outpace systems whose output is increasingly hard to trace.

Several interviewees argue the burden should not rest on individual makers alone. Muralist Terence Tan, known as Ubin Artist, frames the deeper worry: when an image is assembled from precedent, how much of a person remains in it?

Why it matters for Singapore: The city-state courts both AI development and creative media, ambitions that collide in studios where most practitioners freelance alone. Whether consent, disclosure and licensing norms firm up here decides if local illustrators keep charging for original work — or become raw material for models at their doorstep.