Singapore’s AI image boom moves from experimentation to execution

Singapore is fast emerging as one of Asia’s strongest growth markets for ChatGPT Images 2.0, OpenAI’s latest image generation model that can create, refine and iterate visuals from simple prompts, as both consumers and marketers begin integrating the tool into everyday workflows.

According to OpenAI, usage in Singapore jumped more than 80% week-on-week, alongside a more than 90% increase month-on-month, making it one of the region’s top-performing markets behind Taiwan.

The surge is not only driven by existing users. In the past week alone, 50% of image generation activity in Singapore came from first-time users, pointing to strong early curiosity coupled with growing practical adoption. Singapore joins Japan, Thailand and Taiwan as key high-growth markets for the tool across Asia.

Don't miss: ChatGPT Go rolls out globally, ads to be tested on free and Go tiers

The momentum comes as competition in the generative visuals space heats up. ByteDance has been developing Goku AI, an AI-driven video creation framework that transforms text, images and motion signals into fully synthesised video content. Built on a multimodal architecture, it combines inputs such as static images, video and audio cues to generate lifelike human animations, reducing the need for traditional filming and editing workflows.

Adobe is also doubling down on the space with Adobe Firefly, its generative AI suite embedded across Creative Cloud apps. Positioned as a commercially safe model trained on licensed content, Firefly is designed to help creatives generate images, text effects and design assets directly within tools like Photoshop and Illustrator, targeting enterprise and professional users looking for production-ready outputs.

In conversation with MARKETING-INTERACTIVE, Jennifer Lien, head of marketing, APAC, at OpenAI, said the pace of adoption of ChatGPT Images 2.0 signals a broader shift in how users are engaging with AI-generated visuals, moving from experimentation to utility.

“Every major technology shift starts with experimentation. People first ask, ‘Can this do something surprising?’ Then eventually the question becomes, ‘Can this help me work better?’,” she said, adding that image generation is now clearly entering that second phase.

Unlike earlier tools, ChatGPT Images 2.0 introduces what Lien described as “thinking capabilities”, allowing it to search for real-time information, generate multiple outputs from a single prompt, and refine results. This has expanded its role from simple image creation to something more embedded in marketing and communication workflows.

That shift is already reflected in how users are engaging with the tool. Early trends across Singapore and the wider region range from colour analysis and hairstyle simulations to experimenting with makeup looks. At the same time, users are creating product mockups and campaign concepts, blurring the line between casual exploration and commercial application.

Lien said this behaviour is beginning to influence how brands think about creativity and engagement. “We’re seeing people naturally use Images 2.0 in highly visual, identity-driven ways. I believe those same behaviours will increasingly shape how brands think about personalisation and creative engagement,” she said.

In marketing teams, the impact is becoming more tangible. What previously took days of production can now be explored in hours, whether that is testing multiple creative directions, visualising campaign ideas or prototyping packaging concepts.

Lien noted that the tool is also helping brands localise content at scale, particularly in a region as diverse as Asia Pacific, where campaigns often need to adapt across languages and cultural contexts. Improvements in multilingual understanding and non-Latin text rendering are making it easier to produce visuals that feel locally relevant without the same level of production effort.

At the same time, always-on content demands are pushing social teams to produce more with fewer resources. Lien said AI image generation is helping to reduce that burden while still allowing for creative experimentation.

Closing the gap between capability and usage 

Despite the rapid uptake, she noted that a gap still exists between what AI tools are capable of and how widely those capabilities are being used. OpenAI’s research shows that power users are tapping into significantly more advanced features than average users, highlighting the role of experimentation and familiarity.

“With image generation, it makes advanced AI feel much more intuitive and accessible. You don’t need to be technical to start communicating ideas visually,” she said, pointing to the influx of new users in Singapore and other markets as a sign that the barrier to entry is lowering.

Brands, too, are beginning to move beyond using AI purely for internal ideation. Lien said improvements in text rendering, layout consistency and multilingual outputs are making AI-generated visuals increasingly viable for actual campaign development.

“Historically, many image generation tools were great for inspiration, but struggled with text rendering, consistency, or multilingual outputs,” she said. “With Images 2.0, we’re seeing meaningful improvements, which makes the outputs far more usable within actual campaign development workflows.”

Still, she stressed that AI is not replacing creative teams. Instead, its biggest impact will be on compressing the operational layers of the creative process, from early-stage visualisation and mockups to localisation and iterative revisions. This reduces the friction between idea and execution, allowing teams to move faster and explore more directions. Lien added: 

The value isn’t just speed. It’s reducing blank-page starts, helping teams iterate more fluidly, and identifying workflow bottlenecks.

For marketers, the opportunity lies in how the technology is approached. Lien said the teams seeing the strongest results are those treating AI as a collaborator rather than a replacement, using it to accelerate ideation, test creative directions quickly and build fluency across organisations.

As experimentation continues to grow, she added that everyday users may offer the clearest signal of where the space is headed next.

“When everyday users start adapting a tool creatively on their own, it’s usually a sign that something bigger is changing in how people create and communicate,” she said.

That shift also brings new expectations around transparency. A recent study by YouGov and Meltwater found that 84% of Singaporeans believe AI-generated content should be clearly labelled, with 49% saying their trust in a brand would decrease if such use is not disclosed.

The “Trust in the age of generative AI” report, which surveyed nearly 10,000 consumers globally, also highlighted lingering concerns despite growing acceptance. While 55% of Singaporeans said they are excited about AI’s future, 83% expressed concern about its increasing role in daily life.

Misinformation topped the list of worries, with 74% citing risks around fake news or scams, followed by concerns over misleading content (68%), difficulty identifying AI-generated material (64%), and potential misuse of personal data (58%).

Related articles: 
AI momentum builds across Southeast Asia, but gaps persist  
Brands struggle with AI disclosure as usage surges across marketing  
AI ads outperform humans – but only when they don’t look like AI