by

August 23, 2026

15 minutes

Read Time

As generative AI ads flood Singapore market, experts say poorly made ones could backfire on brands, agencies

Creative agencies offering services focusing on AI content told CNA that demand for their services increased sharply in the past six months.

New: You can now listen to articles.

This audio is generated by an AI tool.

Ang Hwee Min

Read a summary of this article on FAST.

Get bite-sized news via a new
cards interface. Give it a try.

Click here to return to FAST
Tap here to return to FAST

FAST

SINGAPORE: Advertisements made with generative artificial intelligence have been flooding the Singapore market in recent months as clients seek to show that they are at the forefront of the technology, marketing industry experts said.

But there are risks involved, as poorly made content is more likely to turn consumers away instead, they added.

The spread of generative AI in the advertising industry is undeniable – clients who use the technology to market their services or products span the gamut – from small- and medium-size enterprises to government agencies.

Songs, films and marketing materials for this year’s National Day celebrations incorporated generative AI. Local brand Mr Bean has introduced its own virtual idol. A scroll on online shopping platforms like Lazada and Shopee turns up AI livestreams from popular household brands like Dyson and Walch.

Guess Word
Crack the word, one row at a time

Buzzword
Create words using the given letters

Mini Sudoku
Tiny puzzle, mighty brain teaser

Mini Crossword
Small grid, big challenge

Word Search
Spot as many words as you can

Show Less

In the past six months, the demand for creative ad agencies that offer services focusing on AI content has shot up sharply, several studios said.

One Singapore advertising agency that specialises in AI-generated content, AI Studio, told CNA it saw slow returns at first, but enquiries have grown by four to five times compared with the period from late 2025 to April 2026.

“When we first started, we thought that mostly small companies would use it, we thought that our crowd would be the SMEs,” said the agency’s founder Carol Tan, 36.

“But I was very wrong because actually most of our clients are enterprise clients. Enterprise clients are the ones who want to be at the forefront of AI,” she said, adding that the agency currently works with multimillion-dollar companies, including ones that are listed.

When Ms Tan, who used to run a traditional marketing and web development agency, first began specialising in AI services about one-and-a-half years ago, a lot of businesses were scared for people to know that they used AI.

“Now actually it’s the other way round. People want to be known as cutting edge.”

On the other hand, there is a fair share of naysayers complaining about or mocking ads that look like they were too obviously made with generative AI.

A quick search on Reddit turns up many threads of online commenters trying to spot the use of generative AI and criticising brands for putting out these ads.

Chairman of the Advertising Standards Authority of Singapore (ASAS) Bryan Tan said it has received two pieces of feedback on the use of generative AI in advertisements in the first half of 2026.

While the level of feedback ASAS has received is about the same as in 2025, the authorities are still concerned if advertisers use generative AI to mislead consumers, do not consider whether their AI use could mislead consumers, or do not vet the generated advertisements for misleading statements and effects, he added.

“Disclosure of the use of AI alone does not whitewash this issue,” said Mr Tan.

Already, half the advertisers out there use generative AI in making video ads, he said earlier this year, pointing to a report by the Interactive Advertising Bureau.

Consumers are concerned about brands using generative AI to create misleading ads, he had said, noting that the feedback it received in 2025 spanned various industries – from interior design and investments to beauty and health.

“Indiscriminate and vague” disclosures of AI use are not useful, he said in April. Instead, ASAS called on marketers and industry bodies to develop guides on how to ethically use generative AI for scenarios unique to their sector.

LOSING CREDIBILITY

When brands use generative AI in their ads, this can signal that they put less effort into the product and its marketing efforts, which can cause them to lose credibility with consumers, experts said.

“Mediocre advertising tends to irritate consumers, irrespective of whether AI was used or not,” said Singapore Management University’s (SMU) Professor of marketing Sabine Benoit.

“But using AI can signal that the company has not put much effort into the product, and that can come across as negligible. This is, of course, the opposite of the signals the organisations want to send.”

While consumers are unlikely to judge a company negatively just because it uses AI, problems arise when the execution feels careless, generic or misleading, she added.

Since generative AI tools are still relatively new, many companies are still learning how to use them well, which is why some “fairly mediocre” work is still emerging, said Prof Benoit.

“Ultimately, using AI will not necessarily damage a brand. Using it poorly can,” she said.

Using generative AI, especially virtual influencers, may be too uncanny for some consumers’ comfort, said Associate Professor Ang Swee Hoon from the National University of Singapore’s (NUS) department of marketing.

“It gives them the creeps,” she said, adding that this causes more mistrust than curiosity where AI influencers are concerned.

The audience may also feel that with the money spent on advertising to get them to buy products, the brand should at least not rely on AI so that what the audience sees is the “truth” or “reality”, and this builds credibility, she added.

Not all publicity is good publicity, experts stressed. The buzz generated online can escalate wildly, and if the buzz is negative, the ad would have generated a negative image for the brand, said Assoc Prof Ang.

In some cases, the ad may have to be taken down. This means the campaign did not achieve its original objectives, she added.

Nanyang Technological University’s (NTU) Assistant Professor Looi Jie Min, who researches digital advertising, said people perceive creativity as a core human trait that stems from personal experience, emotions and cultural understanding.

“I would say that users expect this kind of human involvement for creative fields like advertising and marketing,” she added.

“So when they encounter AI trying to be in these creative fields, they might have a kind of negative reaction because they believe that AI should not be a replacement for human creativity.”

IMPROVED EFFICIENCY AND SCALE

Despite the possible controversy, the demand for AI-generated ads continues to grow. Brands and agencies said using the technology has three main benefits – speed, scale and the ability to demonstrate relevance to the latest trends.

Companies that need to quickly scale ads and marketing materials across many different countries or markets can use generative AI to adapt one set of materials, AI Studio’s founder Ms Tan said, adding that the team can turn a project around in 24 to 48 hours.

“If let’s say they did give it to a traditional agency, it will cost them a lot more to be able to shoot in multiple countries, multiple languages. But with AI, the scale is limitless.”

How comfortable they are with using generative AI also differs from brand to brand. For example, some companies are more cautious about being caught using generative AI, so they only use it for product videos, while other companies don’t mind using it for videos with human models, she added.

Using generative AI also does not necessarily lower costs, said Ms Tan, noting that some brands are afraid of being criticised for raising prices while using AI.

The cost depends on the client’s needs. For example, an SME that might need a very simple one-minute video could pay about S$1,000, while a big MNC that wants a one-minute video with complex, frame-by-frame editing could pay about S$20,000, she added.

Some of the celebrity AI avatars managed by Idolive. (Image: Idolive)

Founder of IdoLive Kevin Zhang, 56, said with online shopping livestreams becoming prevalent, many brands have started to use AI avatars for their livestreams on platforms like Shopee and TikTok Shop.

The business manages the use of AI avatars of celebrities and also provides regular AI avatars for videos and livestreams, which is the bigger portion of its business, he added.

Demand was poor at first, since online shopping platforms used to shut down AI livestreams within about one to two hours, he said, adding that they started accepting such services at the end of 2025.

Since the start of 2026, enquiries for AI livestreaming avatars have increased by 20 to 30 per cent each month, he noted.

“Nowadays it’s so difficult to find a livestreaming host. In Singapore, if you want to get an ordinary host, it’s S$3,000 to S$4,000 for one month’s salary and they can do about four to five hours a day for you,” said Mr Zhang, whose company works with brands like Walch, Dyson and Acer.

Both ordinary human hosts and AI livestream hosts typically see about 10 viewers at any time during their livestream, he added.

But at S$458 per month, an avatar can livestream nonstop for 24 hours a day, said Mr Zhang.

Not only is the cost reduction significant, but an AI avatar can livestream nonstop, which means they bring in more sales since they can sell products online for longer, he stressed.

For the brand to reap the benefits, it has to be relatively well-known and already have a certain number of followers on these online shopping platforms, said Mr Zhang.

This is because consumers already know what products they are looking for and at what price, so it does not matter to them if the livestreamer is a human or an AI avatar, he added.

Agoda’s country director for Singapore Pranjal Kalra said the travel platform used generative AI to create a recent local campaign.

“Without the help of generative AI, executing this across multiple scenarios and formats would have been costly and time-consuming,” she added.

Travel content often needs to be adapted across languages, markets, formats and channels, said Ms Kalra.

Generative AI helps Agoda’s creative teams bring localised content to market faster. For example, using generative AI in video production has allowed Agoda to increase the number of adaptations from 64 to 720 per month, she said.

The time and cost savings differ by the task, and how much human refinement is needed, she added.

“Beyond speed and cost, the practical benefit is faster iteration and the ability to create, localise, and test more creative options efficiently.”

If AI does not improve the work or cannot be used responsibly, it should not be used, she said, adding that it takes user feedback seriously.

A screenshot from a video advertisement from Agoda’s campaign, which was created with generative AI. (Image: YouTube/Agoda)

Ms Priyadharshini Seevaratnam, director of brand and marketing at Circles.Life, said the telco has used AI in its advertising and marketing workflows since 2025.

It uses generative AI tools in copywriting, image creation and video production where appropriate, she told CNA.

Doing so has improved efficiency, she added. Tasks that used to span hours over several production steps can now be completed in a fraction of the time.

The company also reduced its reliance on stock image and stock footage subscriptions, as well as the associated licensing costs, which led to tangible cost savings, said Ms Seevaratnam.

But AI is not an end-to-end solution and every asset made with AI goes through the same quality control process as other campaign materials.

“Our marketing, brand and creative teams review it for accuracy, brand consistency, creative and technical quality, audience suitability, and copyright and usage compliance before final approval and publication,” she added.

AI also does not make final creative decisions or publish content independently, said the telco’s brand and marketing director.

The team also recognises that audiences have differing perspectives on using AI in creative work – to some people, the content can feel impersonal or overly machine-generated, she said.

This is one reason why Circles.Life’s mascot was developed in-house in both 2D and 3D, and not generated by AI, said Ms Seevaratnam.

The team is continuously evaluating whether AI is improving its creative work and serving its customers and brand, she added.

“We would scale back further if AI-assisted work no longer met our quality or brand standards, introduced factual or cultural inaccuracies, or if customer feedback showed that it was reducing relevance or trust.”

Images from PUB’s November 2025 campaign. (Images: Facebook/PUB)

National water agency PUB uses AI to support the development of advertising and marketing materials where appropriate, it said in response to CNA’s queries.

For example, the visuals for PUB’s flood resilience campaign in November 2025 were created using a combination of real actors, photography, stock images and AI-generated elements such as the water and various background props to depict scenes of flooding.

“We take a considered approach to using AI in our creative process. All materials, regardless of the tools used, are carefully reviewed before publication to ensure our communications are accurate and effective, and resonate with our audiences,” said PUB.

Ms Fazilla Nordin, director of corporate communications and strategic partnerships at PropNex Realty, addressed the company’s decision to use generative AI in creating the animated visuals featured in its commercial that played at the National Day Parade.

Many of the scenes originated from actual video footage featuring real people and formed part of the material used in creating the final campaign, she added.

The decision to use generative AI was “primarily a creative one” rather than a cost-reduction exercise, said Ms Fazilla, adding that the creative team was involved throughout.

“This project was therefore also an opportunity for us to experiment, learn and better understand both the possibilities and limitations of generative AI,” she added.

WHO SHOULD AVOID USING GENERATIVE AI

Mr Charlie Tan, the 28-year-old founder of AI agency Touchmkt, sometimes advises clients who are looking to build trust among their consumers or are in high-trust industries against using generative AI.

For example, a tuition centre making a video of customer testimonials should use real parents, children and teachers speaking in the video, to show that it is “legit”, he added.

“But then again, it really depends … What difference does it make if the teacher approves the AI reading a script, versus the actual teacher reading it themselves? We assume the voice and the actions are almost one-to-one, there’s actually not much difference in that sense.”

Larger brands using generative AI despite having the resources to produce ads traditionally represent a recent trend, said NTU’s Asst Prof Looi, noting that this is where most of the backlash comes from.

Research indicates that the backlash is pronounced not just for larger brands, but specifically for luxury brands, she added.

Such brands should be a lot more cautious before jumping onto the AI bandwagon because consumers will see their AI-generated ads as lacking effort and thus perceive the brand as less authentic, said Asst Prof Looi.

A luxury brand is selling an aspirational lifestyle and authenticity is somewhat tied to the high quality of its products, she added.

“If you want to get a consumer to spend big bucks to buy from a luxury brand, and yet your storytelling lacks the effort, why should I be buying from a brand that doesn’t even want to put in the work to get my money?”

Dr Jun Yu, an assistant professor with NUS’ department of communications and new media, noted that a small business might be forgiven for opting to use AI because it lacks the budget, but audiences would perceive that large corporations are using AI by choice.

If a global brand that has spent decades telling its audience how much it cares and can evidently afford to pay people chooses to use AI in its ads, people might think the company has decided its relationship with the audience is no longer worth the expense, he added.

People will start to wonder what else has been quietly automated – customer service, quality checks, the handling of complaints, and so on, said Dr Yu.

“It is not just about a single (instance of) AI use for ad-making, but whether the willingness to spend effort and resources are still there,” said the sociologist.

Dr Yu also pointed out a moral dimension to using generative AI to make ads and marketing materials. More people now know that such technologies were trained on the work of human artists, photographers, writers and other creators without their consent, and the same technologies threaten their livelihoods, he said.

“So a synthetic ad becomes more than just a marketing choice, rather a visible symbol of broader concerns about fairness, inequality and the social costs of AI,” said Dr Yu.

Source: CNA/hw(nj)

Inbox