When we shop online, we just want to know that what we’re seeing is what we’ll get. AI could help us make smarter decisions – or turn us away from buying entirely.

Originally published in print and online by The Straits Times on 24th September 2026

Originally published in print and online by The Straits Times on 24th September 2026

If scrolling through social media and online shopping platforms lately has you feeling like you’ve wandered into the uncanny valley – or worse, been “catfished” into buying something that turns out to be very different from what you thought you were getting – you’re not alone.

Some sellers on Shopee and TikTok Shop recently found AI-generated videos promoting their products with warped book covers, calendars shown at wildly incorrect sizes and, in one particularly bizarre case, a novel about generational trauma and abuse being marketed as a “fun” and “wholesome adventure” for children.

The AI-generated videos were not made by the sellers or with their approval. Even so, the question remains: Has artificial intelligence blurred the line between clever marketing and outright misrepresentation? 

Seeing isn’t believing

Businesses have always made their products look appealing, whether through flattering photography or carefully crafted descriptions. Food advertising has exercised creative licence with visual embellishments – with sandwiches filled with sponges to look fuller, and cereal in a bowl of glue to arrest soggy-looking grains. Yet, as consumers, we know that product images are staged and meant to present the product in the best possible light. 

Generative AI changes that. Often, it goes further than making a product look better. It turns it into something that doesn’t even actually exist. A book jacket could be shown in a vastly different colour or style, and a book made to look as though it belongs to an entirely different genre. 

This is becoming a bigger issue across many e-commerce platforms. Besides Shopee and TikTok Shop, fashion platforms such as Shein are increasingly using AI-generated models and imagery. Amazon, meanwhile, has tools that can conjure up product images and styles from little more than a text prompt. 

And, unlike some of the more obvious Photoshop fails we have grown accustomed to spotting, this kind of manipulation can be impossible to identify. An AI-generated image can have all the visual hallmarks we have learnt to trust in a photograph, like convincing lighting, shadows and textures.

In short, it’s becoming increasingly harder to tell whether what you’re seeing is what you get when the product reaches you. 

Online shopping already asks consumers to take a leap of faith and buy an item they have not seen, touched or tried on in real life. Generative AI applied on such platforms turns consumers off and erodes trust when people finally receive their purchases, only to find they had bought something entirely different from what was presented.

The price to pay

From a consumer perspective, there’s an irony here. When we shop online, we ultimately want to see a representation of a product that’s as close to reality as possible. We also seek out affirmation of quality – unbiased reviews or real-life customer photos – when presented with artificial visuals. 

But reading every single product review before deciding on a purchase takes a mental toll. More likely, users will leave the shopping app and head to social media channels for a chance to see what the product actually looks like.

There is something reassuring about seeing a dress worn to an actual dinner, or a piece of furniture photographed in someone’s badly lit living room.

It’s evidence that the thing exists somewhere outside the beautifully rendered world of the product page.

But that cycle isn’t sustainable for time-strapped consumers, and platforms that enable indiscriminate use of generative AI could pay a price for this. 

If shoppers increasingly have to leave a platform to figure out what a product really looks like, the polished images meant to make a sale may be doing the opposite. 

One misleading image may be forgotten, but repeated experiences can make consumers more sceptical of what they see, more likely to look elsewhere for reassurance, or simply more willing to take their money elsewhere. 

In fact, the Shopee and TikTok Shop incidents reflect how quickly that problem can land back on businesses. The sellers ended up reporting the AI-generated videos, while others restricted how their products could be promoted, citing fears that their brand reputation would take a hit. 

When a business leans too heavily on AI-generated marketing content, shoppers may begin to see it not as innovative, but untrustworthy – and cutting corners at their expense.

AI done right

That doesn’t mean AI has no place in shopping. Take Fit:match, a fashion technology company that uses body scanning and AI to recommend clothing based on a shopper’s measurements and fit preferences. Rather than generating a more flattering version of a dress, the technology helps answer one of the most frustrating questions in online shopping: Will this actually fit me? 

Home Depot takes a similarly practical approach with its AI shopping assistant, Magic Apron, which can help customers find products, compare options and work through DIY projects. If you are standing in front of a broken tap wondering whether you need a new tap, a wrench or to simply call a plumber, an AI assistant can be considerably more useful than another product video.

AI can also make shopping more accessible. Be My Eyes, for example, uses AI to help people who are blind or have low vision make sense of what is in front of them. Its Be My AI assistant can describe images, read information and answer questions about a user’s surroundings, making it easier to navigate everything from a product label to the contents of a shopping basket. 

What all these applications have in common is that they use AI to make the shopper better informed. If AI can help people compare options or anticipate the things they may otherwise discover only after the package arrives, that seems like a pretty good use of the technology.

But if it exists mainly to make the product look better than reality, we are back where we started.

There’s a reason why “expectation versus reality” videos are trending on social media. In these videos, shoppers film the moment they hold up what arrived against the image that sold it to them. Often, they look nothing like their listings. 

For all the excitement around making shopping more immersive, frictionless and increasingly artificial, what consumers ultimately want is surprisingly old-fashioned: to know what they’re buying.