Skip to content

A picture is worth a thousand words, except with AI

Images generated in Adobe Firefly to emulate a Type G BS 1363 socket with London in the background.

I’ve been working on a website about electrical outlets in my spare time. Exciting stuff, I know. Mostly, the site serves as a sandbox for me where I can experiment with AI and SEO tools simultaneously. If it generates any revenue in the long run, that’ll simply be a plus.

Most recently, I’ve compiled a series of articles about the electrical outlets in specific countries. To that end, I’d like to include more than just a wall of AI-generated text. Videos, images and any multimedia boosts user engagement and signals to Google that you’re putting together a whole website.

But here’s where it’s tricky: Simply embedding Creative Commons images of various outlet and plug types is both visually unappealing and won’t snag me any design awards. While the site is fairly utilitarian in nature, it should still be original and have some aesthetic thought. I also don’t really want to pay someone on Upwork to snap a photo of their outlet with their phone.

AI can seemingly solve this problem; plug in a CC image, give a prompt to zhuzh things up, et voilà. You have art.

Except it doesn’t quite work so well.

AI image tools struggle with technical prompts

I’ve been playing around with a variety of free AI tools, so I haven’t opted to pay for Midjourney. So far, I’ve tested Krea AI and Adobe Firefly, since the former is free and the latter is part of my Adobe subscription.

I tried plugging in a CC image of a Type G outlet, which is used in the U.K., into Firefly. I then wanted to play up the British motif, so I asked Firefly to create an image of a Type G socket with London as the background. The problem is, the socket doesn’t have the correct number of holes.

The image on the right is the CC image I used for reference in Firefly. The image on the left is one of the four created by Firefly based on the prompt “British Standard type G electrical outlet with london in background.” Initially, I tried using “British Standard BS 1363 electrical outlet with london in background,” but Firefly told me, “One or more words may not meet User Guidelines and were removed.” BS, indeed.

Regardless, I’ve tried a few different prompts in hopes of generating a usable image that reflects what a Type G socket looks like to no avail. While the images are aesthetically pleasing, they’re inaccurate. The same happened when trying to generate one for the plug, rather than the outlet.

Those familiar with Type G plugs and sockets will note that none of these images reflect the familiar, chonky, three-pin plug. In other words, none of the images are accurate. I’ve played around with prompts and the level of AI, but thus far, have not netted a better result.

I’ll keep playing around with AI image tools, and they seem on the path to solve problems like the one I have. But for today, designers, illustrators and photographers can probably rest easy.

Leave a Reply

Your email address will not be published. Required fields are marked *