The action figure trend is the latest way people are misusing the power of AI – and I wish I could stop doing it

With the regrettable Studio Ghbli AI trend in our rearview mirror, we've stumbled right into our next unfortunate AI trend: Action Figures of ourselves.

To be clear, I've made one of myself with ChatGPT 4o's new native image capabilities because how could I not? As you might expect, ChatGPT's work is exceptional. I fed OpenAI's GPT 4o model a full-length photo of me and used this prompt:

Create an action figure of the person in this photo. It must be in its packaging and with things that represent his interests like computers, robots, cameras, smartphones.

The initial results were good, though I neglected to tell ChatGPT to use my name on the packaging. The platform also stumbled a bit when I asked for adjustments to the figure, with each iteration looking less and less like me.

It doesn't matter. I did it, and so have thousands of others. At least we're not, as with the Ghibli and Muppets AI memes, stealing someone else's intellectual property (IP) to create new and uncanny images.

I can understand why everyone is doing it. First of all, these images look just like real action figure packaging. The addition of interest accessories and, though I didn't ask for it, an optional head, is perfect.

There is a proportional relationship between this quality and how quickly these AI image trends spread. The generative images are so good that as soon as they started to appear on social media, others started investigating how to make one for themselves.

AI Action Figures in packaging are so popular that there are, unsurprisingly, YouTube tutorials. That's how I figured out how to do it. I found a Spanish-language one created about a week ago. The translation gave me just enough detail to know how to form the proper action figure prompt.

The true cost of AI memes

This is all good fun, but there are concerns.

First of all, AI image generation is not without cost. Sure, there’s the price of a ChatGPT Plus membership (around $20 / £16 / AU$30 a month), although you can generate around three images a day on the free tier, depending on current demand. Perhaps more importantly, there’s the cost of AI models like 4o.

A Queens University Library report claims, "Artificial Intelligence models consume an enormous amount of water and emit large amounts of carbon in their production, training, operation, and maintenance." Another Cornell University study calls out AI's growing freshwater use footprint, claiming "training the GPT-3 language model in Microsoft's state-of-the-art U.S. data centers can directly evaporate 700,000 liters of clean freshwater."

If you don't think these AI trends and the memes they spawn are attracting wide use, stressing the system, and possibly eating natural resources, just look at OpenAI CEO Sam Altman's comments.

We have a joke in my house that every time we create one of these AI memes, it kills a tree. That's hyperbole, of course, but it's safe to say that AI content generation is not without costs, and perhaps we should be thinking about it and using it differently,

That is the other question. With so many consumers primarily using AI for trend image generation, are they missing the point?

The growing intelligence of AI means that someday in the not-too-distant future it could be as smart or smarter than us (Artificial General intelligence could arrive as soon as next year). Meanwhile, we'll all be using it to create funny movie posters and then shocked when AI takes our jobs.

The way into AI and perhaps keeping control of the narrative is by using AI as a practical tool. It's not as much fun, but ChatGPT, Gemini, Copilot, Apple Intelligence, and soon Alexa+, all feature myriad tools for improving things we do every day, including writing, creating outlines, polishing presentations, and summarizing long articles. Yes, they can create complete images and short videos, but tools like Adobe Firefly are more helpful for adjusting existing ones in more subtle ways, like removing backgrounds or distracting people and objects.

There are people using AIs like ChatGPT 4o for these tasks but those workaday activities are drowned out by a flood of AI trend images and memes like the ones above of my life-like action figure in plastic packaging.

It is kind of cool, right?

Bonus

If you drop that action figure image into Sora, you can craft a prompt that depicts it (you) being removed from your packaging.

You might also like

• I discovered a surprising difference between DeepSeek and ChatGPT search capabilities
• I tried ChatGPT's new sarcastic voice, and it made me hate Mondays even more
• ChatGPT just wrote the most beautiful short story, and I wonder what I'm even doing here
• ChatGPT 4o image generation is so good we will never be able to trust iPhone renders again