Generative AI advertising has become hard to avoid in Singapore this year, marketing industry professionals say, as businesses of every size race to show they are keeping pace with the technology. The clients turning to AI made content span the full range of the market, from small local shops to government agencies, and the pitch is largely the same, a desire to look current rather than dated. That rush has a downside. Several marketing academics and the country's advertising watchdog caution that content produced carelessly is more likely to push consumers away than draw them in.
AI has already reached the national stage
The scale of the shift is visible in some of the most visible campaigns of the year. Songs, films, and marketing material for this year's National Day celebrations incorporated generative AI. Local brand Mr Bean rolled out its own virtual idol. Scroll through online shopping platforms such as Lazada and Shopee and AI hosted livestreams from household names including Dyson and Walch are easy to find. What was once a novelty has become a routine part of how brands show up online.
Agencies say enquiries have multiplied
Carol Tan, 36, founder of the Singapore agency AI Studio, said her business grew slowly at first but enquiries have climbed four to five times over compared with the period between late 2025 and April 2026. She expected small and medium enterprises to make up most of her client base when she pivoted from a traditional marketing and web development agency about a year and a half ago. Instead, most of her clients turned out to be large, sometimes publicly listed, companies eager to be seen as forward thinking.
The attitude toward disclosure has flipped just as sharply, Ms Tan said. Businesses that once wanted to hide their use of AI now want it known, treating the technology as a badge of being cutting edge rather than something to downplay.
Online, the reaction is far less friendly
Not every audience shares that enthusiasm. Threads on Reddit regularly pick apart ads that look obviously machine made, with commenters calling brands out by name. Bryan Tan, chairman of the Advertising Standards Authority of Singapore, said the regulator received two pieces of feedback specifically about generative AI in ads during the first half of 2026, roughly matching the level seen in 2025. His office remains watchful of advertisers who use AI to mislead shoppers, who fail to consider whether their AI output could mislead, or who skip vetting generated ads for problematic claims. Simply telling consumers that an ad used AI, he said, does not resolve those concerns on its own.
Mr Tan pointed earlier this year to an Interactive Advertising Bureau report suggesting that half of all advertisers already use generative AI somewhere in their video ad production, with complaints in 2025 touching industries as varied as interior design, investment products, beauty, and health. Rather than vague AI disclaimers, he has pushed marketers and trade bodies to build sector specific guidance on ethical use.
Why a mediocre AI ad costs more than it saves
Sabine Benoit, professor of marketing at Singapore Management University, said the underlying problem is rarely AI itself. Weak advertising annoys audiences whether or not a machine was involved, she said, but AI adds a specific risk, it can read as a signal that a company did not put in the effort, the opposite of what any brand wants to communicate. Consumers are unlikely to penalize a brand simply for using the technology, she added. The trouble starts when the execution feels careless, generic, or misleading, which is common while many companies are still learning to use the tools well.
Ang Swee Hoon, associate professor of marketing at the National University of Singapore, pointed to a separate risk with AI generated spokespeople. Virtual influencers can feel too close to human without quite clearing the bar, she said, describing the effect as one that gives viewers the creeps and breeds mistrust rather than curiosity. Audiences who are effectively paying for advertising with their attention, she added, tend to want the reassurance that what they are watching is real.
Assistant Professor Looi Jie Min of Nanyang Technological University, who studies digital advertising, said much of the resistance comes down to how people define creativity itself, as something rooted in lived experience, emotion, and cultural context. Audiences expect a human hand in creative fields like advertising, she said, and tend to react negatively when AI appears to be standing in for that human judgment rather than assisting it.
Not all attention is good attention, either. A poorly received AI ad can generate exactly the kind of viral backlash a brand was hoping to avoid, sometimes forcing the campaign offline before it meets any of its original goals.
The case marketers keep making anyway
Despite the risk of backlash, demand keeps climbing, and the people selling AI services point to three consistent advantages, speed, scale, and the ability to look current. Ms Tan said her team can turn projects around in 24 to 48 hours, work that would take far longer and cost far more through a traditional agency shooting across multiple countries and languages. Pricing swings widely depending on complexity. A simple one minute video for a small business might run around S$1,000, she said, while a major multinational client asking for intricate, frame by frame editing could pay closer to S$20,000. AI does not automatically mean cheaper, either, since some clients worry about being criticized for raising prices while leaning on automated production.
AI hosts are reshaping livestream shopping
The shift is especially visible in livestream commerce. Kevin Zhang, 56, founder of IdoLive, which manages AI avatars for brands including Walch and Acer, said enquiries for AI livestream hosts have risen 20 to 30 percent every month since the start of 2026. A human host in Singapore typically costs S$3,000 to S$4,000 a month for four to five hours of streaming a day, and tends to draw around ten viewers at a time, roughly the same audience an AI avatar attracts. An AI host, by contrast, costs about S$458 a month and can stream around the clock, which Mr Zhang said translates into more selling time and more sales, provided the brand already has enough of a following that shoppers already know what they came to buy.
Bigger brands describe a more cautious approach
Larger companies say they are treating AI as one tool among many rather than a wholesale replacement for creative teams. Agoda's Singapore country director Pranjal Kalra said generative AI helped the travel platform run a recent local campaign across far more formats than would otherwise have been affordable, since travel marketing often needs adapting across languages, markets, and channels. He said AI video production let Agoda scale its adaptations from 64 to 720 a month, describing the real advantage as faster iteration and testing rather than cost cutting alone. If a use of AI does not clearly improve the work or cannot be handled responsibly, he said, the company will not use it.
At telco Circles.Life, brand and marketing director Priyadharshini Seevaratnam said the company has used generative tools for copywriting, image creation, and video production since 2025, cutting production time and reducing its reliance on paid stock image and footage libraries. Every AI assisted asset still passes through the same review as any other campaign material, she said, checked for accuracy, brand consistency, creative and technical quality, audience fit, and rights compliance before anything is published. AI does not make final creative calls or publish anything on its own, she added, and the company deliberately built its mascot in house in both 2D and 3D rather than generating it, aware that some audiences find AI made characters impersonal. The team said it would pull back further if AI ever hurt quality, introduced factual or cultural errors, or eroded customer trust.
The real dividing line
Across the interviews, a consistent theme emerges. Whether a brand benefits from generative AI or gets burned by it appears to depend less on the technology and more on the discipline applied around it, the same editorial standards, fact checking, and brand judgment that any other piece of advertising would need to earn an audience's trust.
