I recently experimented with Midjourney to generate an image of two hikers. One was wearing pilot-frame sunglasses, which I liked.
A day or two later, I walked into a Costco and saw a pair of Hugo Boss sunglasses on a special display that looked like what I saw on the AI hikers. They were on sale for a great price. I tried them on, looked at myself in the mirror, and put them back to get what was on my shopping list.
Well, a few minutes later, I went back and bought them. I got a new pair of sunglasses at a good price, and Costco got my money. Win, win?
Product placement
Product placement isn’t a new marketing strategy. Companies have been paying for characters to use their products since radio was the entertainment source. Sometimes it is subtle, while other times, it is obvious or clunky.
I’m not sure if product placement in AI-generated images is a fantasy or a nightmare, but it is an opportunity.
Trade-offs
Currently, many AI image generators charge a subscription fee. That’s understandable; the tech and the infrastructure they run on are not cheap.
It is not a matter of if but when a service will partner with a company to offer a discount in exchange for something. Perhaps you can use the tool for free or at a discounted rate if you’re willing to have people or things in the images you generate wear a hat with a company’s logo, for instance.
At some point during the image generation process, the user will be able to click or tap on the hat, which will be linked to the sponsor’s website. Martech practitioners will certainly help with such campaigns.
I’m not sure if that’s what I would want most of the time, but the discounted subscription rate does have some appeal.
Dig deeper: The new frontier of visual content: A marketer’s guide to AI
The SEO analogy
A more subtle approach to product placement would follow an SEO approach, and that seems more interesting.
It also doesn’t seem unlikely that people will (or already are?) strategize how to make specific images stand out in the vast swaths of data that AI models pull from to generate imagery. Instead of focusing on search terms, keywords in the prompt would take precedence.
If a company made sunglasses, what if an AI service heavily relied upon images of the eyewear company’s products to generate images? Instead of having a noticeable product placement scenario, it is much more subtle.
This is where my recent experience would fall. I have no idea of the brand or model of the sunglasses that the AI-generated hiker was wearing (it was all concocted, after all). But they looked great, and close enough to the Hugo Boss ones that Costco was showcasing when I walked in.
Granted, AI does not have a great track record of displaying things well. For instance, AI tools distort human hands a lot. There’s also the possibility the product will be included in an image of someone doing something violent or otherwise controversial. Brand safety around this will be tough.
I am sure that product designers, intellectual property lawyers and reputation managers may fret over AI tools hallucinating and garbling what a product looks like. However, AI tool users will prompt tools to generate images of their products, whether they like them or not.
This is where SEO-like tactics will come in. Experts will emerge providing strategies on structuring image metadata, using HTML tags for publicly available images, configuring and exposing DAMs to datasets and so on.
AI platforms will likely monetize their data by, for instance, providing a way for people to see what terms people use in prompts. Then image asset owners can optimize their metadata and catalogs accordingly. Let’s also think about analytics folks for a minute. How would they assign attribution for this?
Image searches
We can certainly take this a step further. For instance, Microsoft Bing and Google provide tools for people to isolate something in an image to conduct an image-based search. Someone may see a photo of a celebrity and want to buy something they see in it. They can use that image search to find that item or a dupe (how the cool kids these days refer to knockoffs).
If you were a martech practitioner at an apparel company, how would you optimize your product images and pages to show up in such image search results?
I have tried this with AI-generated images. Although the sunglasses I saw were conjured up by AI imagination, Google Lens provides results for items similar in appearance. This is another way that AI services can monetize their data.
Dig deeper: 4 AI tools to supercharge visual content creation
Fantasy or nightmare
I’m not looking forward to such a future, but does anyone doubt this is all plausible?
Such possibilities are important to consider regardless of how they’ll unfold. The possibility of product placement highlights the importance of having a martech strategy. Thus, it is a great reminder that we all should strive to manage, categorize, and structure assets and data in a way that is accessible and easy for both humans and algorithms to understand.
Contributing authors are invited to create content for MarTech and are chosen for their expertise and contribution to the martech community. Our contributors work under the oversight of the editorial staff and contributions are checked for quality and relevance to our readers. The opinions they express are their own.
