Getty Images/iStockphoto

How Amazon AI image generation could help advertisers

The tech giant joins others using generative AI technology to cater to advertisers and help small businesses scale their advertising through personalization.

Advertisers are among the business groups that stand to benefit the most from the early wave of generative AI-supported software products.

Among other vendors targeting ad creators, Amazon is out this week with a new AI image generation tool in beta for advertisers.

The tool, unveiled Wednesday, enables advertisers to improve ad performance by adding background images to their products, according to Amazon.

For example, instead of an advertiser having its product -- say, a toaster -- against a white background, now, with image generation, the toaster can be placed on a kitchen counter or next to a croissant.

This level of placement increases click-through rates, according to Amazon.

Generative AI saves time

Amazon's new features come amid various generative AI (GenAI)-based tools and products for advertisers released in previous months.

In September, eBay introduced a feature that lets sellers create entire product listings from a photo using a generative AI tool.

Earlier this month, Meta previewed new features available for advertisers in Meta's Ads Manager. YouTube on Oct. 18 introduced Spotlight Moments, a package that enables advertisers and marketers to capitalize on significant events.

All these platforms are using GenAI to cut time out of [making] creatives.
Eric SchmittAnalyst, Gartner

"All these platforms are using GenAI to cut time out of [making] creatives," or creative products such as ads, said Gartner analyst Eric Schmitt.

As for Amazon's image generation product, he added that advertisers seeking to sell items on Amazon alone are the ones the vendor is targeting.

However, advertisers looking to place images on not only Amazon but also other advertising platforms might be better off with some of Adobe's products, Schmitt continued. This is because it doesn't appear that images created with Amazon's image generation tool can be used beyond the Amazon platform.

Amazon's image generation is going to evolve as the tool allows users to create multiple versions of their creative products, he added.

"We will see more versions of creative tailored for different purposes," Schmitt said. "You can see how that capability will become more important over time."

Personalization at scale

It's likely Amazon's customers asked for a tool like this, said Constellation Research analyst Liz Miller.

"A lot of this is being driven by the reality that people understand that there are only so many units of creative you can create in a day," Miller said.

Not only is Amazon helping advertisers save time but it's also using its experience in connecting consumers to the right products to help advertisers with personalized ads at scale, she continued.

"The light has come on; the technology has come to bear where personalization is possible at scale," Miller said.

GenAI changes the game for advertisers because now they can ask the AI system if they are introducing the correct type of advertising that a customer wants to see.

While many large companies have been taking part in personalization, with a tool like image generation, small to midsize businesses can also take advantage of these tools in their advertising strategy, Miller said.

Image generation, however, is challenging due to some of the criticisms that generative AI tools like Stable Diffusion and Dall-E have received since their introduction. For example, some say that when such tools scrape data from the web, they illegally and unethically copy human artists' work. Also, some critics contend that the technology could displace artists' and designers' jobs and reduce their potential income.

Meanwhile, image-generation technology has accelerated in just the past few months, Miller said.

For example, the difference between the first Adobe Firefly image model that came out in March and the recently introduced Firefly Image 2 Model is exponential. Miller added that the level of detail, perspective and color changed greatly from the first to the second.

"We've lived leaps and bounds over the last nine months," she said. "It's going to continue."

Esther Ajao is a TechTarget Editorial news writer covering artificial intelligence software and systems.

Dig Deeper on Enterprise applications of AI

Business Analytics
Data Management