‘Ghibli-inspired’ AI images

OK, an interesting one for a Friday morning.

You’ve probably seen some of the so called ‘Ghibli-inspired’ images flooding the news and social media today. If you’ve not, here’s one of the many stories about it. The basic premise is that ChatGPT is ripping off Studio Ghibli’s style.

I’m sure most of you know this, but Studio Ghibli is a Japanese film studio, responsible for Spirited Away etc.

Let’s step back a bit here though.

Are the comments and approach from OpenAI/Sam Altman tone deaf? Absolutely.

But to me this feels very similar to the debates around AI generated music in that the debate is implying AI is replicating a style, when, frankly, it’s getting no where near it.

So, full disclosure. I’m a massive long-term fan of the work two of Studio Ghibili’s leading directors – Hayao Miyazaki and Isao Takahata. I’d quite happily argue that Princess Mononoke is the greatest film ever made. Is this odd for a middle-aged bloke from Wales? Don’t judge me until you’ve watched a lot of the films!

Let’s start with the basics. There’s no such thing as a Ghibli style. The different directors have hugely different approach to their films, and none of the directors have a consistent style. Like all artists their work evolves over time. If I was to be charitable, and say by Ghibli-style people mean latter day Miyazaki – it still just doesn’t hold for me. The images just look like generic, vaguely Japanese animation inspired images. I feel there’s a bit of an odd cultural undertone here, like saying any vaguely tuneful guitar music sounds like the Beatles or whatever.

I know others disagree with this, and say they images look just like Ghibli images. Maybe ones from Goro Miyazaki (in-joke for Ghibli fans!). Art is subjective. That’s fine.

Why does this matter? It feels the same as the discussion about AI generated music. What’s created is technically music, just as the images are technically animation inspired images. But they just aren’t getting close to the real thing. They might work for fun or anonymous background music, or throwaway images to illustrate a point, but are they really a threat to human artistry? Honestly I really don’t think so.

I’m actually much more optimistic that these AI tools will just be more in the armour for actual artists to help them create fantastic, moving, inspiring work. But the tech companies really need to diversify their employee based away from just techies. Their overall approach and messaging is just wrong at the moment.