How Can Artists Use AI Without Losing the Craft (and How Not To)

The AI sloppification of content has done a good job of obscuring the fact that AI can be genuinely useful for artists, if used in the right ways. Seeing an avalanche of AI-produced content pouring onto feeds, somehow looking scarily good and nightmarishly uncanny at the same time, it is completely understandable that many artists choose to reject AI outright as a kneejerk reaction.

But behind the slopalanche, many artists and studios are quietly finding ways to use AI that improve their output while retaining the craft.

The key is to stop assuming that all AI use looks like the plastic, trope-laden output of someone prompting Midjourney and posting the result straight to Instagram. It also goes further than using ChatGPT to write your captions. The real unlock is subtler than either of those, and it starts with understanding what makes creative work feel like creative work in the first place.

The Process Is the Point

The most important principle when using AI in creative production is this: you should still be working in a way that feels natural for an artist or a team of artists.

Consider what this is not. Using AI-generated images as the basis for AI-generated video. Prompting images is not a natural way to work. Sketching and painting is. The AI-image-to-video pipeline also tends to favour a single person working alone, which strips out the layers of collaboration that are usually baked into any serious production. On a stop motion shoot or a feature film set, dozens of people contribute their expertise at every stage. Each pair of eyes, each set of hands, each round of feedback makes the thing better. When you reduce that to one person typing prompts, something essential is lost.

A common misconception about AI video in general is that you need to start with AI-generated images. In my experience, AI images skip the necessary steps an artist usually goes through to realise their vision. You are jumping straight to the end, cutting off the feedback loop between the artist and the work, between the artist and their collaborators, between the artist's eye and the slow, generative act of making something by hand.

Where the Real Unlock Lives

For certain styles of animation, it can be genuinely transformative to render characters and environments at high resolution using traditional 3D tools, then animate them with an AI video model. The reason is practical: 3D character rigging is technically demanding, especially when you are dealing with complex surfaces like fur. The AI handles the motion while the artists handle everything else.

Myth Studio's work on the Nestle Compleat project is a good example of what this looks like in practice. The TV commercial was created with an entirely traditional process up to the point of animation. It started with hand-drawn storyboards and concept art. An almost impressionist colour script captured the mood and palette. The team modelled detailed 3D environments, placing every prop and texturing every surface. The mother and daughter possum characters were modelled in 3D. We then created high-resolution renders of each scene and animated them with AI video models.

From the Inchstones production: colour script, 3D environment, and character model, all created by hand before AI animation.

So if the animation is ultimately handled by AI, why go through the painstaking work of creating every preliminary step by hand?

Drawing and painting remain the best way to get creative ideas out of your head and into the world. Building a 3D environment is like constructing a film set. The artist can move a camera around to find the right angle. Play with focal lengths. Adjust bloom and lighting. This iterative feedback between the artist's eye and the emerging work is precisely the thing that gets cauterised when you skip the process. It is also what separates considered production from AI slop. On the Compleat project, we had a large, multidisciplined team, each contributing their specific skills, guided by a strong creative direction.

And what of the AI video part itself? Interestingly, when you reduce AI's role to animating high-quality assets rather than generating everything from scratch, the work becomes something closer to directing. The right AI-animator, someone with strong taste and a considered system for prompting, can coax remarkable performances from rendered characters. The generated performances are a little unruly by nature, but when tamed and carefully selected, they can work very well for the right project.

Other Ways Artists Are Using AI

The Through Line

Every useful application of AI in creative work shares the same underlying principle: the artist is still working in a way that feels natural and collaborative. The AI brings speed and occasionally a lateral perspective, but the creative decisions, the taste, the craft, those remain human. The moment the tool starts replacing the process rather than augmenting it, you are back in slop territory.

The slopalanche is not going away. But for artists and studios willing to look past it, there are real, considered ways to bring these tools into the work without losing what makes the work worth doing in the first place.